{"id":1188,"date":"2026-09-20T22:36:45","date_gmt":"2026-09-20T22:36:45","guid":{"rendered":"https:\/\/virtualserversvps.com\/blog\/?p=1188"},"modified":"2026-09-20T22:36:45","modified_gmt":"2026-09-20T22:36:45","slug":"cgroups-v2-cap-memory-oversubscribed-vps","status":"publish","type":"post","link":"https:\/\/virtualserversvps.com\/blog\/cgroups-v2-cap-memory-oversubscribed-vps\/","title":{"rendered":"Capping Memory with cgroups v2 on an Oversubscribed VPS"},"content":{"rendered":"<li><strong>Forgetting swap.<\/strong> <code>memory.swap.max=0<\/code> on a database is often correct, but on a stateless web worker it turns manageable pressure into a hard kill.<\/li><li><strong>Ignoring nested cgroups.<\/strong> A limit on a parent slice and a child unit both apply; the more restrictive one wins. Check the full path with <code>systemd-cgls<\/code>.<\/li><li><strong>Trusting old guides.<\/strong> Anything referencing <code>memory.limit_in_bytes<\/code>, <code>tasks<\/code> or <code>cgroup.clone_children<\/code> is cgroups v1 and will not work on a modern VPS.<\/li><\/ul>\n\n\n\n<h2 class=\"wp-block-heading\">Where this fits<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">cgroup memory limits do not make a workload fit in less RAM. They make the failure mode predictable: instead of a global OOM that takes out MariaDB because it happened to be the largest RSS, you get a scoped kill of the process that actually misbehaved, and a counter in <code>memory.events<\/code> telling you it happened. On an oversubscribed VPS that is the difference between a degraded service and a dead one. Pair it with correct application-level sizing \u2014 see our <a href=\"https:\/\/virtualserversvps.com\/blog\/tuning-php-fpm-process-pools-by-workload\">PHP-FPM pool sizing guide<\/a> \u2014 and you have a system that fails locally instead of globally.<\/p>\n\n<!-- \/wp:list-item --><li><strong>Setting only <code>memory.max<\/code>.<\/strong> Without <code>memory.high<\/code> you lose the graceful throttle and go straight to a kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Forgetting swap.<\/strong> <code>memory.swap.max=0<\/code> on a database is often correct, but on a stateless web worker it turns manageable pressure into a hard kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Ignoring nested cgroups.<\/strong> A limit on a parent slice and a child unit both apply; the more restrictive one wins. Check the full path with <code>systemd-cgls<\/code>.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Trusting old guides.<\/strong> Anything referencing <code>memory.limit_in_bytes<\/code>, <code>tasks<\/code> or <code>cgroup.clone_children<\/code> is cgroups v1 and will not work on a modern VPS.<\/li><!-- \/wp:list-item --><\/ul>\n<!-- \/wp:list -->\n\n<!-- wp:heading {\"level\":2} -->\n<h2>Where this fits<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>cgroup memory limits do not make a workload fit in less RAM. They make the failure mode predictable: instead of a global OOM that takes out MariaDB because it happened to be the largest RSS, you get a scoped kill of the process that actually misbehaved, and a counter in <code>memory.events<\/code> telling you it happened. On an oversubscribed VPS that is the difference between a degraded service and a dead one. Pair it with correct application-level sizing \u2014 see our <a href=\"https:\/\/virtualserversvps.com\/blog\/tuning-php-fpm-process-pools-by-workload\">PHP-FPM pool sizing guide<\/a> \u2014 and you have a system that fails locally instead of globally.<\/p>\n<!-- \/wp:paragraph -->\n<!-- \/wp:list --><!-- \/wp:post-content --><!-- \/wp:list-item --><!-- wp:list-item --><li><strong>Setting only <code>memory.max<\/code>.<\/strong> Without <code>memory.high<\/code> you lose the graceful throttle and go straight to a kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Forgetting swap.<\/strong> <code>memory.swap.max=0<\/code> on a database is often correct, but on a stateless web worker it turns manageable pressure into a hard kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Ignoring nested cgroups.<\/strong> A limit on a parent slice and a child unit both apply; the more restrictive one wins. Check the full path with <code>systemd-cgls<\/code>.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Trusting old guides.<\/strong> Anything referencing <code>memory.limit_in_bytes<\/code>, <code>tasks<\/code> or <code>cgroup.clone_children<\/code> is cgroups v1 and will not work on a modern VPS.<\/li><!-- \/wp:list-item --><\/ul>\n<!-- \/wp:list -->\n\n<!-- wp:heading {\"level\":2} -->\n<h2>Where this fits<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>cgroup memory limits do not make a workload fit in less RAM. They make the failure mode predictable: instead of a global OOM that takes out MariaDB because it happened to be the largest RSS, you get a scoped kill of the process that actually misbehaved, and a counter in <code>memory.events<\/code> telling you it happened. On an oversubscribed VPS that is the difference between a degraded service and a dead one. Pair it with correct application-level sizing \u2014 see our <a href=\"https:\/\/virtualserversvps.com\/blog\/tuning-php-fpm-process-pools-by-workload\">PHP-FPM pool sizing guide<\/a> \u2014 and you have a system that fails locally instead of globally.<\/p>\n<!-- \/wp:paragraph -->\n<!-- \/wp:list --><!-- \/wp:post-content --><!-- \/wp:list-item -->\n\n<!-- wp:list -->\n<ul><!-- wp:list-item --><li><strong>Assuming <code>memory.max<\/code> caps RSS.<\/strong> It caps the cgroup&#8217;s total memory including page cache charged to it. A process reading a 300&nbsp;MB file may hit its limit with a small RSS.<\/li><!-- \/wp:list-item --><!-- wp:list-item --><li><strong>Setting only <code>memory.max<\/code>.<\/strong> Without <code>memory.high<\/code> you lose the graceful throttle and go straight to a kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Forgetting swap.<\/strong> <code>memory.swap.max=0<\/code> on a database is often correct, but on a stateless web worker it turns manageable pressure into a hard kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Ignoring nested cgroups.<\/strong> A limit on a parent slice and a child unit both apply; the more restrictive one wins. Check the full path with <code>systemd-cgls<\/code>.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Trusting old guides.<\/strong> Anything referencing <code>memory.limit_in_bytes<\/code>, <code>tasks<\/code> or <code>cgroup.clone_children<\/code> is cgroups v1 and will not work on a modern VPS.<\/li><!-- \/wp:list-item --><\/ul>\n<!-- \/wp:list -->\n\n<!-- wp:heading {\"level\":2} -->\n<h2>Where this fits<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>cgroup memory limits do not make a workload fit in less RAM. They make the failure mode predictable: instead of a global OOM that takes out MariaDB because it happened to be the largest RSS, you get a scoped kill of the process that actually misbehaved, and a counter in <code>memory.events<\/code> telling you it happened. On an oversubscribed VPS that is the difference between a degraded service and a dead one. Pair it with correct application-level sizing \u2014 see our <a href=\"https:\/\/virtualserversvps.com\/blog\/tuning-php-fpm-process-pools-by-workload\">PHP-FPM pool sizing guide<\/a> \u2014 and you have a system that fails locally instead of globally.<\/p>\n<!-- \/wp:paragraph -->\n<!-- \/wp:list --><!-- \/wp:post-content --><!-- \/wp:list --><li><strong>Forgetting swap.<\/strong> <code>memory.swap.max=0<\/code> on a database is often correct, but on a stateless web worker it turns manageable pressure into a hard kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Ignoring nested cgroups.<\/strong> A limit on a parent slice and a child unit both apply; the more restrictive one wins. Check the full path with <code>systemd-cgls<\/code>.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Trusting old guides.<\/strong> Anything referencing <code>memory.limit_in_bytes<\/code>, <code>tasks<\/code> or <code>cgroup.clone_children<\/code> is cgroups v1 and will not work on a modern VPS.<\/li><!-- \/wp:list-item --><\/ul>\n<!-- \/wp:list-item -->\n\n<!-- wp:heading {\"level\":2} -->\n<h2>Where this fits<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>cgroup memory limits do not make a workload fit in less RAM. They make the failure mode predictable: instead of a global OOM that takes out MariaDB because it happened to be the largest RSS, you get a scoped kill of the process that actually misbehaved, and a counter in <code>memory.events<\/code> telling you it happened. On an oversubscribed VPS that is the difference between a degraded service and a dead one. Pair it with correct application-level sizing \u2014 see our <a href=\"https:\/\/virtualserversvps.com\/blog\/tuning-php-fpm-process-pools-by-workload\">PHP-FPM pool sizing guide<\/a> \u2014 and you have a system that fails locally instead of globally.<\/p>\n<!-- \/wp:paragraph -->\n<!-- \/wp:list-item -->\n\n<!-- wp:list -->\n<ul><!-- wp:list-item --><li><strong>Assuming <code>memory.max<\/code> caps RSS.<\/strong> It caps the cgroup&#8217;s total memory including page cache charged to it. A process reading a 300&nbsp;MB file may hit its limit with a small RSS.<\/li><!-- \/wp:list-item --><!-- wp:list-item --><li><strong>Setting only <code>memory.max<\/code>.<\/strong> Without <code>memory.high<\/code> you lose the graceful throttle and go straight to a kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Forgetting swap.<\/strong> <code>memory.swap.max=0<\/code> on a database is often correct, but on a stateless web worker it turns manageable pressure into a hard kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Ignoring nested cgroups.<\/strong> A limit on a parent slice and a child unit both apply; the more restrictive one wins. Check the full path with <code>systemd-cgls<\/code>.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Trusting old guides.<\/strong> Anything referencing <code>memory.limit_in_bytes<\/code>, <code>tasks<\/code> or <code>cgroup.clone_children<\/code> is cgroups v1 and will not work on a modern VPS.<\/li><!-- \/wp:list-item --><\/ul>\n<!-- \/wp:list -->\n\n<!-- wp:heading {\"level\":2} -->\n<h2>Where this fits<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>cgroup memory limits do not make a workload fit in less RAM. They make the failure mode predictable: instead of a global OOM that takes out MariaDB because it happened to be the largest RSS, you get a scoped kill of the process that actually misbehaved, and a counter in <code>memory.events<\/code> telling you it happened. On an oversubscribed VPS that is the difference between a degraded service and a dead one. Pair it with correct application-level sizing \u2014 see our <a href=\"https:\/\/virtualserversvps.com\/blog\/tuning-php-fpm-process-pools-by-workload\">PHP-FPM pool sizing guide<\/a> \u2014 and you have a system that fails locally instead of globally.<\/p>\n<!-- \/wp:paragraph -->\n<!-- \/wp:list --><!-- \/wp:post-content --><!-- \/wp:list-item --><!-- wp:list-item --><li><strong>Forgetting swap.<\/strong> <code>memory.swap.max=0<\/code> on a database is often correct, but on a stateless web worker it turns manageable pressure into a hard kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Ignoring nested cgroups.<\/strong> A limit on a parent slice and a child unit both apply; the more restrictive one wins. Check the full path with <code>systemd-cgls<\/code>.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Trusting old guides.<\/strong> Anything referencing <code>memory.limit_in_bytes<\/code>, <code>tasks<\/code> or <code>cgroup.clone_children<\/code> is cgroups v1 and will not work on a modern VPS.<\/li><!-- \/wp:list-item --><\/ul>\n<!-- \/wp:list -->\n\n<!-- wp:heading {\"level\":2} -->\n<h2>Where this fits<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>cgroup memory limits do not make a workload fit in less RAM. They make the failure mode predictable: instead of a global OOM that takes out MariaDB because it happened to be the largest RSS, you get a scoped kill of the process that actually misbehaved, and a counter in <code>memory.events<\/code> telling you it happened. On an oversubscribed VPS that is the difference between a degraded service and a dead one. Pair it with correct application-level sizing \u2014 see our <a href=\"https:\/\/virtualserversvps.com\/blog\/tuning-php-fpm-process-pools-by-workload\">PHP-FPM pool sizing guide<\/a> \u2014 and you have a system that fails locally instead of globally.<\/p>\n<!-- \/wp:paragraph -->\n<!-- \/wp:list-item -->\n\n<!-- wp:list -->\n<ul><!-- wp:list-item --><li><strong>Assuming <code>memory.max<\/code> caps RSS.<\/strong> It caps the cgroup&#8217;s total memory including page cache charged to it. A process reading a 300&nbsp;MB file may hit its limit with a small RSS.<\/li><!-- \/wp:list-item --><!-- wp:list-item --><li><strong>Setting only <code>memory.max<\/code>.<\/strong> Without <code>memory.high<\/code> you lose the graceful throttle and go straight to a kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Forgetting swap.<\/strong> <code>memory.swap.max=0<\/code> on a database is often correct, but on a stateless web worker it turns manageable pressure into a hard kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Ignoring nested cgroups.<\/strong> A limit on a parent slice and a child unit both apply; the more restrictive one wins. Check the full path with <code>systemd-cgls<\/code>.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Trusting old guides.<\/strong> Anything referencing <code>memory.limit_in_bytes<\/code>, <code>tasks<\/code> or <code>cgroup.clone_children<\/code> is cgroups v1 and will not work on a modern VPS.<\/li><!-- \/wp:list-item --><\/ul>\n<!-- \/wp:list -->\n\n<!-- wp:heading {\"level\":2} -->\n<h2>Where this fits<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>cgroup memory limits do not make a workload fit in less RAM. They make the failure mode predictable: instead of a global OOM that takes out MariaDB because it happened to be the largest RSS, you get a scoped kill of the process that actually misbehaved, and a counter in <code>memory.events<\/code> telling you it happened. On an oversubscribed VPS that is the difference between a degraded service and a dead one. Pair it with correct application-level sizing \u2014 see our <a href=\"https:\/\/virtualserversvps.com\/blog\/tuning-php-fpm-process-pools-by-workload\">PHP-FPM pool sizing guide<\/a> \u2014 and you have a system that fails locally instead of globally.<\/p>\n<!-- \/wp:paragraph -->\n<!-- \/wp:list --><!-- \/wp:post-content --><!-- \/wp:list-item --><li><strong>Setting only <code>memory.max<\/code>.<\/strong> Without <code>memory.high<\/code> you lose the graceful throttle and go straight to a kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Forgetting swap.<\/strong> <code>memory.swap.max=0<\/code> on a database is often correct, but on a stateless web worker it turns manageable pressure into a hard kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Ignoring nested cgroups.<\/strong> A limit on a parent slice and a child unit both apply; the more restrictive one wins. Check the full path with <code>systemd-cgls<\/code>.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Trusting old guides.<\/strong> Anything referencing <code>memory.limit_in_bytes<\/code>, <code>tasks<\/code> or <code>cgroup.clone_children<\/code> is cgroups v1 and will not work on a modern VPS.<\/li><!-- \/wp:list-item --><\/ul>\n<!-- \/wp:list -->\n\n<!-- wp:heading {\"level\":2} -->\n<h2>Where this fits<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>cgroup memory limits do not make a workload fit in less RAM. They make the failure mode predictable: instead of a global OOM that takes out MariaDB because it happened to be the largest RSS, you get a scoped kill of the process that actually misbehaved, and a counter in <code>memory.events<\/code> telling you it happened. On an oversubscribed VPS that is the difference between a degraded service and a dead one. Pair it with correct application-level sizing \u2014 see our <a href=\"https:\/\/virtualserversvps.com\/blog\/tuning-php-fpm-process-pools-by-workload\">PHP-FPM pool sizing guide<\/a> \u2014 and you have a system that fails locally instead of globally.<\/p>\n<!-- \/wp:paragraph -->\n<!-- \/wp:list-item -->\n\n<!-- wp:list -->\n<ul><!-- wp:list-item --><li><strong>Assuming <code>memory.max<\/code> caps RSS.<\/strong> It caps the cgroup&#8217;s total memory including page cache charged to it. A process reading a 300&nbsp;MB file may hit its limit with a small RSS.<\/li><!-- \/wp:list-item --><!-- wp:list-item --><li><strong>Setting only <code>memory.max<\/code>.<\/strong> Without <code>memory.high<\/code> you lose the graceful throttle and go straight to a kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Forgetting swap.<\/strong> <code>memory.swap.max=0<\/code> on a database is often correct, but on a stateless web worker it turns manageable pressure into a hard kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Ignoring nested cgroups.<\/strong> A limit on a parent slice and a child unit both apply; the more restrictive one wins. Check the full path with <code>systemd-cgls<\/code>.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Trusting old guides.<\/strong> Anything referencing <code>memory.limit_in_bytes<\/code>, <code>tasks<\/code> or <code>cgroup.clone_children<\/code> is cgroups v1 and will not work on a modern VPS.<\/li><!-- \/wp:list-item --><\/ul>\n<!-- \/wp:list -->\n\n<!-- wp:heading {\"level\":2} -->\n<h2>Where this fits<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>cgroup memory limits do not make a workload fit in less RAM. They make the failure mode predictable: instead of a global OOM that takes out MariaDB because it happened to be the largest RSS, you get a scoped kill of the process that actually misbehaved, and a counter in <code>memory.events<\/code> telling you it happened. On an oversubscribed VPS that is the difference between a degraded service and a dead one. Pair it with correct application-level sizing \u2014 see our <a href=\"https:\/\/virtualserversvps.com\/blog\/tuning-php-fpm-process-pools-by-workload\">PHP-FPM pool sizing guide<\/a> \u2014 and you have a system that fails locally instead of globally.<\/p>\n<!-- \/wp:paragraph -->\n<!-- \/wp:list --><!-- \/wp:post-content --><!-- \/wp:list-item --><!-- wp:list-item --><li><strong>Setting only <code>memory.max<\/code>.<\/strong> Without <code>memory.high<\/code> you lose the graceful throttle and go straight to a kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Forgetting swap.<\/strong> <code>memory.swap.max=0<\/code> on a database is often correct, but on a stateless web worker it turns manageable pressure into a hard kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Ignoring nested cgroups.<\/strong> A limit on a parent slice and a child unit both apply; the more restrictive one wins. Check the full path with <code>systemd-cgls<\/code>.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Trusting old guides.<\/strong> Anything referencing <code>memory.limit_in_bytes<\/code>, <code>tasks<\/code> or <code>cgroup.clone_children<\/code> is cgroups v1 and will not work on a modern VPS.<\/li><!-- \/wp:list-item --><\/ul>\n<!-- \/wp:list -->\n\n<!-- wp:heading {\"level\":2} -->\n<h2>Where this fits<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>cgroup memory limits do not make a workload fit in less RAM. They make the failure mode predictable: instead of a global OOM that takes out MariaDB because it happened to be the largest RSS, you get a scoped kill of the process that actually misbehaved, and a counter in <code>memory.events<\/code> telling you it happened. On an oversubscribed VPS that is the difference between a degraded service and a dead one. Pair it with correct application-level sizing \u2014 see our <a href=\"https:\/\/virtualserversvps.com\/blog\/tuning-php-fpm-process-pools-by-workload\">PHP-FPM pool sizing guide<\/a> \u2014 and you have a system that fails locally instead of globally.<\/p>\n<!-- \/wp:paragraph -->\n<!-- \/wp:list-item -->\n\n<!-- wp:list -->\n<ul><!-- wp:list-item --><li><strong>Assuming <code>memory.max<\/code> caps RSS.<\/strong> It caps the cgroup&#8217;s total memory including page cache charged to it. A process reading a 300&nbsp;MB file may hit its limit with a small RSS.<\/li><!-- \/wp:list-item --><!-- wp:list-item --><li><strong>Setting only <code>memory.max<\/code>.<\/strong> Without <code>memory.high<\/code> you lose the graceful throttle and go straight to a kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Forgetting swap.<\/strong> <code>memory.swap.max=0<\/code> on a database is often correct, but on a stateless web worker it turns manageable pressure into a hard kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Ignoring nested cgroups.<\/strong> A limit on a parent slice and a child unit both apply; the more restrictive one wins. Check the full path with <code>systemd-cgls<\/code>.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Trusting old guides.<\/strong> Anything referencing <code>memory.limit_in_bytes<\/code>, <code>tasks<\/code> or <code>cgroup.clone_children<\/code> is cgroups v1 and will not work on a modern VPS.<\/li><!-- \/wp:list-item --><\/ul>\n<!-- \/wp:list -->\n\n<!-- wp:heading {\"level\":2} -->\n<h2>Where this fits<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>cgroup memory limits do not make a workload fit in less RAM. They make the failure mode predictable: instead of a global OOM that takes out MariaDB because it happened to be the largest RSS, you get a scoped kill of the process that actually misbehaved, and a counter in <code>memory.events<\/code> telling you it happened. On an oversubscribed VPS that is the difference between a degraded service and a dead one. Pair it with correct application-level sizing \u2014 see our <a href=\"https:\/\/virtualserversvps.com\/blog\/tuning-php-fpm-process-pools-by-workload\">PHP-FPM pool sizing guide<\/a> \u2014 and you have a system that fails locally instead of globally.<\/p>\n<!-- \/wp:paragraph -->\n<!-- \/wp:list --><!-- \/wp:post-content --><!-- \/wp:list-item --><li><strong>Forgetting swap.<\/strong> <code>memory.swap.max=0<\/code> on a database is often correct, but on a stateless web worker it turns manageable pressure into a hard kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Ignoring nested cgroups.<\/strong> A limit on a parent slice and a child unit both apply; the more restrictive one wins. Check the full path with <code>systemd-cgls<\/code>.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Trusting old guides.<\/strong> Anything referencing <code>memory.limit_in_bytes<\/code>, <code>tasks<\/code> or <code>cgroup.clone_children<\/code> is cgroups v1 and will not work on a modern VPS.<\/li><!-- \/wp:list-item --><\/ul>\n<!-- \/wp:list-item -->\n\n<!-- wp:heading {\"level\":2} -->\n<h2>Where this fits<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>cgroup memory limits do not make a workload fit in less RAM. They make the failure mode predictable: instead of a global OOM that takes out MariaDB because it happened to be the largest RSS, you get a scoped kill of the process that actually misbehaved, and a counter in <code>memory.events<\/code> telling you it happened. On an oversubscribed VPS that is the difference between a degraded service and a dead one. Pair it with correct application-level sizing \u2014 see our <a href=\"https:\/\/virtualserversvps.com\/blog\/tuning-php-fpm-process-pools-by-workload\">PHP-FPM pool sizing guide<\/a> \u2014 and you have a system that fails locally instead of globally.<\/p>\n<!-- \/wp:paragraph -->\n<!-- \/wp:list-item --><!-- wp:list-item --><li><strong>Setting only <code>memory.max<\/code>.<\/strong> Without <code>memory.high<\/code> you lose the graceful throttle and go straight to a kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Forgetting swap.<\/strong> <code>memory.swap.max=0<\/code> on a database is often correct, but on a stateless web worker it turns manageable pressure into a hard kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Ignoring nested cgroups.<\/strong> A limit on a parent slice and a child unit both apply; the more restrictive one wins. Check the full path with <code>systemd-cgls<\/code>.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Trusting old guides.<\/strong> Anything referencing <code>memory.limit_in_bytes<\/code>, <code>tasks<\/code> or <code>cgroup.clone_children<\/code> is cgroups v1 and will not work on a modern VPS.<\/li><!-- \/wp:list-item --><\/ul>\n<!-- \/wp:list -->\n\n<!-- wp:heading {\"level\":2} -->\n<h2>Where this fits<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>cgroup memory limits do not make a workload fit in less RAM. They make the failure mode predictable: instead of a global OOM that takes out MariaDB because it happened to be the largest RSS, you get a scoped kill of the process that actually misbehaved, and a counter in <code>memory.events<\/code> telling you it happened. On an oversubscribed VPS that is the difference between a degraded service and a dead one. Pair it with correct application-level sizing \u2014 see our <a href=\"https:\/\/virtualserversvps.com\/blog\/tuning-php-fpm-process-pools-by-workload\">PHP-FPM pool sizing guide<\/a> \u2014 and you have a system that fails locally instead of globally.<\/p>\n<!-- \/wp:paragraph -->\n<!-- \/wp:list-item -->\n\n<!-- wp:list -->\n<ul><!-- wp:list-item --><li><strong>Assuming <code>memory.max<\/code> caps RSS.<\/strong> It caps the cgroup&#8217;s total memory including page cache charged to it. A process reading a 300&nbsp;MB file may hit its limit with a small RSS.<\/li><!-- \/wp:list-item --><!-- wp:list-item --><li><strong>Setting only <code>memory.max<\/code>.<\/strong> Without <code>memory.high<\/code> you lose the graceful throttle and go straight to a kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Forgetting swap.<\/strong> <code>memory.swap.max=0<\/code> on a database is often correct, but on a stateless web worker it turns manageable pressure into a hard kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Ignoring nested cgroups.<\/strong> A limit on a parent slice and a child unit both apply; the more restrictive one wins. Check the full path with <code>systemd-cgls<\/code>.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Trusting old guides.<\/strong> Anything referencing <code>memory.limit_in_bytes<\/code>, <code>tasks<\/code> or <code>cgroup.clone_children<\/code> is cgroups v1 and will not work on a modern VPS.<\/li><!-- \/wp:list-item --><\/ul>\n<!-- \/wp:list -->\n\n<!-- wp:heading {\"level\":2} -->\n<h2>Where this fits<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>cgroup memory limits do not make a workload fit in less RAM. They make the failure mode predictable: instead of a global OOM that takes out MariaDB because it happened to be the largest RSS, you get a scoped kill of the process that actually misbehaved, and a counter in <code>memory.events<\/code> telling you it happened. On an oversubscribed VPS that is the difference between a degraded service and a dead one. Pair it with correct application-level sizing \u2014 see our <a href=\"https:\/\/virtualserversvps.com\/blog\/tuning-php-fpm-process-pools-by-workload\">PHP-FPM pool sizing guide<\/a> \u2014 and you have a system that fails locally instead of globally.<\/p>\n<!-- \/wp:paragraph -->\n<!-- \/wp:list --><!-- \/wp:post-content --><!-- \/wp:list-item --><!-- wp:list-item --><li><strong>Forgetting swap.<\/strong> <code>memory.swap.max=0<\/code> on a database is often correct, but on a stateless web worker it turns manageable pressure into a hard kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Ignoring nested cgroups.<\/strong> A limit on a parent slice and a child unit both apply; the more restrictive one wins. Check the full path with <code>systemd-cgls<\/code>.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Trusting old guides.<\/strong> Anything referencing <code>memory.limit_in_bytes<\/code>, <code>tasks<\/code> or <code>cgroup.clone_children<\/code> is cgroups v1 and will not work on a modern VPS.<\/li><!-- \/wp:list-item --><\/ul>\n<!-- \/wp:list -->\n\n<!-- wp:heading {\"level\":2} -->\n<h2>Where this fits<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>cgroup memory limits do not make a workload fit in less RAM. They make the failure mode predictable: instead of a global OOM that takes out MariaDB because it happened to be the largest RSS, you get a scoped kill of the process that actually misbehaved, and a counter in <code>memory.events<\/code> telling you it happened. On an oversubscribed VPS that is the difference between a degraded service and a dead one. Pair it with correct application-level sizing \u2014 see our <a href=\"https:\/\/virtualserversvps.com\/blog\/tuning-php-fpm-process-pools-by-workload\">PHP-FPM pool sizing guide<\/a> \u2014 and you have a system that fails locally instead of globally.<\/p>\n<!-- \/wp:paragraph -->\n<!-- \/wp:list-item --><!-- wp:list-item --><li><strong>Setting only <code>memory.max<\/code>.<\/strong> Without <code>memory.high<\/code> you lose the graceful throttle and go straight to a kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Forgetting swap.<\/strong> <code>memory.swap.max=0<\/code> on a database is often correct, but on a stateless web worker it turns manageable pressure into a hard kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Ignoring nested cgroups.<\/strong> A limit on a parent slice and a child unit both apply; the more restrictive one wins. Check the full path with <code>systemd-cgls<\/code>.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Trusting old guides.<\/strong> Anything referencing <code>memory.limit_in_bytes<\/code>, <code>tasks<\/code> or <code>cgroup.clone_children<\/code> is cgroups v1 and will not work on a modern VPS.<\/li><!-- \/wp:list-item --><\/ul>\n<!-- \/wp:list -->\n\n<!-- wp:heading {\"level\":2} -->\n<h2>Where this fits<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>cgroup memory limits do not make a workload fit in less RAM. They make the failure mode predictable: instead of a global OOM that takes out MariaDB because it happened to be the largest RSS, you get a scoped kill of the process that actually misbehaved, and a counter in <code>memory.events<\/code> telling you it happened. On an oversubscribed VPS that is the difference between a degraded service and a dead one. Pair it with correct application-level sizing \u2014 see our <a href=\"https:\/\/virtualserversvps.com\/blog\/tuning-php-fpm-process-pools-by-workload\">PHP-FPM pool sizing guide<\/a> \u2014 and you have a system that fails locally instead of globally.<\/p>\n<!-- \/wp:paragraph -->\n<!-- \/wp:list-item -->\n\n<!-- wp:list -->\n<ul><!-- wp:list-item --><li><strong>Assuming <code>memory.max<\/code> caps RSS.<\/strong> It caps the cgroup&#8217;s total memory including page cache charged to it. A process reading a 300&nbsp;MB file may hit its limit with a small RSS.<\/li><!-- \/wp:list-item --><!-- wp:list-item --><li><strong>Setting only <code>memory.max<\/code>.<\/strong> Without <code>memory.high<\/code> you lose the graceful throttle and go straight to a kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Forgetting swap.<\/strong> <code>memory.swap.max=0<\/code> on a database is often correct, but on a stateless web worker it turns manageable pressure into a hard kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Ignoring nested cgroups.<\/strong> A limit on a parent slice and a child unit both apply; the more restrictive one wins. Check the full path with <code>systemd-cgls<\/code>.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Trusting old guides.<\/strong> Anything referencing <code>memory.limit_in_bytes<\/code>, <code>tasks<\/code> or <code>cgroup.clone_children<\/code> is cgroups v1 and will not work on a modern VPS.<\/li><!-- \/wp:list-item --><\/ul>\n<!-- \/wp:list -->\n\n<!-- wp:heading {\"level\":2} -->\n<h2>Where this fits<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>cgroup memory limits do not make a workload fit in less RAM. They make the failure mode predictable: instead of a global OOM that takes out MariaDB because it happened to be the largest RSS, you get a scoped kill of the process that actually misbehaved, and a counter in <code>memory.events<\/code> telling you it happened. On an oversubscribed VPS that is the difference between a degraded service and a dead one. Pair it with correct application-level sizing \u2014 see our <a href=\"https:\/\/virtualserversvps.com\/blog\/tuning-php-fpm-process-pools-by-workload\">PHP-FPM pool sizing guide<\/a> \u2014 and you have a system that fails locally instead of globally.<\/p>\n<!-- \/wp:paragraph -->\n<!-- \/wp:list --><!-- \/wp:post-content --><!-- \/wp:list-item --><li><strong>Setting only <code>memory.max<\/code>.<\/strong> Without <code>memory.high<\/code> you lose the graceful throttle and go straight to a kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Forgetting swap.<\/strong> <code>memory.swap.max=0<\/code> on a database is often correct, but on a stateless web worker it turns manageable pressure into a hard kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Ignoring nested cgroups.<\/strong> A limit on a parent slice and a child unit both apply; the more restrictive one wins. Check the full path with <code>systemd-cgls<\/code>.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Trusting old guides.<\/strong> Anything referencing <code>memory.limit_in_bytes<\/code>, <code>tasks<\/code> or <code>cgroup.clone_children<\/code> is cgroups v1 and will not work on a modern VPS.<\/li><!-- \/wp:list-item --><\/ul>\n<!-- \/wp:list -->\n\n<!-- wp:heading {\"level\":2} -->\n<h2>Where this fits<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>cgroup memory limits do not make a workload fit in less RAM. They make the failure mode predictable: instead of a global OOM that takes out MariaDB because it happened to be the largest RSS, you get a scoped kill of the process that actually misbehaved, and a counter in <code>memory.events<\/code> telling you it happened. On an oversubscribed VPS that is the difference between a degraded service and a dead one. Pair it with correct application-level sizing \u2014 see our <a href=\"https:\/\/virtualserversvps.com\/blog\/tuning-php-fpm-process-pools-by-workload\">PHP-FPM pool sizing guide<\/a> \u2014 and you have a system that fails locally instead of globally.<\/p>\n<!-- \/wp:paragraph -->\n<!-- \/wp:list-item --><!-- wp:list-item --><li><strong>Setting only <code>memory.max<\/code>.<\/strong> Without <code>memory.high<\/code> you lose the graceful throttle and go straight to a kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Forgetting swap.<\/strong> <code>memory.swap.max=0<\/code> on a database is often correct, but on a stateless web worker it turns manageable pressure into a hard kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Ignoring nested cgroups.<\/strong> A limit on a parent slice and a child unit both apply; the more restrictive one wins. Check the full path with <code>systemd-cgls<\/code>.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Trusting old guides.<\/strong> Anything referencing <code>memory.limit_in_bytes<\/code>, <code>tasks<\/code> or <code>cgroup.clone_children<\/code> is cgroups v1 and will not work on a modern VPS.<\/li><!-- \/wp:list-item --><\/ul>\n<!-- \/wp:list -->\n\n<!-- wp:heading {\"level\":2} -->\n<h2>Where this fits<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>cgroup memory limits do not make a workload fit in less RAM. They make the failure mode predictable: instead of a global OOM that takes out MariaDB because it happened to be the largest RSS, you get a scoped kill of the process that actually misbehaved, and a counter in <code>memory.events<\/code> telling you it happened. On an oversubscribed VPS that is the difference between a degraded service and a dead one. Pair it with correct application-level sizing \u2014 see our <a href=\"https:\/\/virtualserversvps.com\/blog\/tuning-php-fpm-process-pools-by-workload\">PHP-FPM pool sizing guide<\/a> \u2014 and you have a system that fails locally instead of globally.<\/p>\n<!-- \/wp:paragraph -->\n<!-- \/wp:list-item -->\n\n<!-- wp:list -->\n<ul><!-- wp:list-item --><li><strong>Assuming <code>memory.max<\/code> caps RSS.<\/strong> It caps the cgroup&#8217;s total memory including page cache charged to it. A process reading a 300&nbsp;MB file may hit its limit with a small RSS.<\/li><!-- \/wp:list-item --><!-- wp:list-item --><li><strong>Setting only <code>memory.max<\/code>.<\/strong> Without <code>memory.high<\/code> you lose the graceful throttle and go straight to a kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Forgetting swap.<\/strong> <code>memory.swap.max=0<\/code> on a database is often correct, but on a stateless web worker it turns manageable pressure into a hard kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Ignoring nested cgroups.<\/strong> A limit on a parent slice and a child unit both apply; the more restrictive one wins. Check the full path with <code>systemd-cgls<\/code>.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Trusting old guides.<\/strong> Anything referencing <code>memory.limit_in_bytes<\/code>, <code>tasks<\/code> or <code>cgroup.clone_children<\/code> is cgroups v1 and will not work on a modern VPS.<\/li><!-- \/wp:list-item --><\/ul>\n<!-- \/wp:list -->\n\n<!-- wp:heading {\"level\":2} -->\n<h2>Where this fits<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>cgroup memory limits do not make a workload fit in less RAM. They make the failure mode predictable: instead of a global OOM that takes out MariaDB because it happened to be the largest RSS, you get a scoped kill of the process that actually misbehaved, and a counter in <code>memory.events<\/code> telling you it happened. On an oversubscribed VPS that is the difference between a degraded service and a dead one. Pair it with correct application-level sizing \u2014 see our <a href=\"https:\/\/virtualserversvps.com\/blog\/tuning-php-fpm-process-pools-by-workload\">PHP-FPM pool sizing guide<\/a> \u2014 and you have a system that fails locally instead of globally.<\/p>\n<!-- \/wp:paragraph -->\n<!-- \/wp:list --><!-- \/wp:post-content --><!-- \/wp:list-item --><li><strong>Forgetting swap.<\/strong> <code>memory.swap.max=0<\/code> on a database is often correct, but on a stateless web worker it turns manageable pressure into a hard kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Ignoring nested cgroups.<\/strong> A limit on a parent slice and a child unit both apply; the more restrictive one wins. Check the full path with <code>systemd-cgls<\/code>.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Trusting old guides.<\/strong> Anything referencing <code>memory.limit_in_bytes<\/code>, <code>tasks<\/code> or <code>cgroup.clone_children<\/code> is cgroups v1 and will not work on a modern VPS.<\/li><!-- \/wp:list-item --><\/ul>\n<!-- \/wp:list-item -->\n\n<!-- wp:heading {\"level\":2} -->\n<h2>Where this fits<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>cgroup memory limits do not make a workload fit in less RAM. They make the failure mode predictable: instead of a global OOM that takes out MariaDB because it happened to be the largest RSS, you get a scoped kill of the process that actually misbehaved, and a counter in <code>memory.events<\/code> telling you it happened. On an oversubscribed VPS that is the difference between a degraded service and a dead one. Pair it with correct application-level sizing \u2014 see our <a href=\"https:\/\/virtualserversvps.com\/blog\/tuning-php-fpm-process-pools-by-workload\">PHP-FPM pool sizing guide<\/a> \u2014 and you have a system that fails locally instead of globally.<\/p>\n<!-- \/wp:paragraph -->\n<!-- \/wp:list-item --><li><strong>Setting only <code>memory.max<\/code>.<\/strong> Without <code>memory.high<\/code> you lose the graceful throttle and go straight to a kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Forgetting swap.<\/strong> <code>memory.swap.max=0<\/code> on a database is often correct, but on a stateless web worker it turns manageable pressure into a hard kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Ignoring nested cgroups.<\/strong> A limit on a parent slice and a child unit both apply; the more restrictive one wins. Check the full path with <code>systemd-cgls<\/code>.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Trusting old guides.<\/strong> Anything referencing <code>memory.limit_in_bytes<\/code>, <code>tasks<\/code> or <code>cgroup.clone_children<\/code> is cgroups v1 and will not work on a modern VPS.<\/li><!-- \/wp:list-item --><\/ul>\n<!-- \/wp:list -->\n\n<!-- wp:heading {\"level\":2} -->\n<h2>Where this fits<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>cgroup memory limits do not make a workload fit in less RAM. They make the failure mode predictable: instead of a global OOM that takes out MariaDB because it happened to be the largest RSS, you get a scoped kill of the process that actually misbehaved, and a counter in <code>memory.events<\/code> telling you it happened. On an oversubscribed VPS that is the difference between a degraded service and a dead one. Pair it with correct application-level sizing \u2014 see our <a href=\"https:\/\/virtualserversvps.com\/blog\/tuning-php-fpm-process-pools-by-workload\">PHP-FPM pool sizing guide<\/a> \u2014 and you have a system that fails locally instead of globally.<\/p>\n<!-- \/wp:paragraph -->\n<!-- \/wp:list-item --><!-- wp:list-item --><li><strong>Setting only <code>memory.max<\/code>.<\/strong> Without <code>memory.high<\/code> you lose the graceful throttle and go straight to a kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Forgetting swap.<\/strong> <code>memory.swap.max=0<\/code> on a database is often correct, but on a stateless web worker it turns manageable pressure into a hard kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Ignoring nested cgroups.<\/strong> A limit on a parent slice and a child unit both apply; the more restrictive one wins. Check the full path with <code>systemd-cgls<\/code>.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Trusting old guides.<\/strong> Anything referencing <code>memory.limit_in_bytes<\/code>, <code>tasks<\/code> or <code>cgroup.clone_children<\/code> is cgroups v1 and will not work on a modern VPS.<\/li><!-- \/wp:list-item --><\/ul>\n<!-- \/wp:list -->\n\n<!-- wp:heading {\"level\":2} -->\n<h2>Where this fits<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>cgroup memory limits do not make a workload fit in less RAM. They make the failure mode predictable: instead of a global OOM that takes out MariaDB because it happened to be the largest RSS, you get a scoped kill of the process that actually misbehaved, and a counter in <code>memory.events<\/code> telling you it happened. On an oversubscribed VPS that is the difference between a degraded service and a dead one. Pair it with correct application-level sizing \u2014 see our <a href=\"https:\/\/virtualserversvps.com\/blog\/tuning-php-fpm-process-pools-by-workload\">PHP-FPM pool sizing guide<\/a> \u2014 and you have a system that fails locally instead of globally.<\/p>\n<!-- \/wp:paragraph -->\n<!-- \/wp:list-item -->\n\n<!-- wp:list -->\n<ul><!-- wp:list-item --><li><strong>Assuming <code>memory.max<\/code> caps RSS.<\/strong> It caps the cgroup&#8217;s total memory including page cache charged to it. A process reading a 300&nbsp;MB file may hit its limit with a small RSS.<\/li><!-- \/wp:list-item --><!-- wp:list-item --><li><strong>Setting only <code>memory.max<\/code>.<\/strong> Without <code>memory.high<\/code> you lose the graceful throttle and go straight to a kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Forgetting swap.<\/strong> <code>memory.swap.max=0<\/code> on a database is often correct, but on a stateless web worker it turns manageable pressure into a hard kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Ignoring nested cgroups.<\/strong> A limit on a parent slice and a child unit both apply; the more restrictive one wins. Check the full path with <code>systemd-cgls<\/code>.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Trusting old guides.<\/strong> Anything referencing <code>memory.limit_in_bytes<\/code>, <code>tasks<\/code> or <code>cgroup.clone_children<\/code> is cgroups v1 and will not work on a modern VPS.<\/li><!-- \/wp:list-item --><\/ul>\n<!-- \/wp:list -->\n\n<!-- wp:heading {\"level\":2} -->\n<h2>Where this fits<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>cgroup memory limits do not make a workload fit in less RAM. They make the failure mode predictable: instead of a global OOM that takes out MariaDB because it happened to be the largest RSS, you get a scoped kill of the process that actually misbehaved, and a counter in <code>memory.events<\/code> telling you it happened. On an oversubscribed VPS that is the difference between a degraded service and a dead one. Pair it with correct application-level sizing \u2014 see our <a href=\"https:\/\/virtualserversvps.com\/blog\/tuning-php-fpm-process-pools-by-workload\">PHP-FPM pool sizing guide<\/a> \u2014 and you have a system that fails locally instead of globally.<\/p>\n<!-- \/wp:paragraph -->\n<!-- \/wp:list --><!-- \/wp:post-content --><!-- \/wp:list-item --><!-- wp:list-item --><li><strong>Forgetting swap.<\/strong> <code>memory.swap.max=0<\/code> on a database is often correct, but on a stateless web worker it turns manageable pressure into a hard kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Ignoring nested cgroups.<\/strong> A limit on a parent slice and a child unit both apply; the more restrictive one wins. Check the full path with <code>systemd-cgls<\/code>.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Trusting old guides.<\/strong> Anything referencing <code>memory.limit_in_bytes<\/code>, <code>tasks<\/code> or <code>cgroup.clone_children<\/code> is cgroups v1 and will not work on a modern VPS.<\/li><!-- \/wp:list-item --><\/ul>\n<!-- \/wp:list -->\n\n<!-- wp:heading {\"level\":2} -->\n<h2>Where this fits<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>cgroup memory limits do not make a workload fit in less RAM. They make the failure mode predictable: instead of a global OOM that takes out MariaDB because it happened to be the largest RSS, you get a scoped kill of the process that actually misbehaved, and a counter in <code>memory.events<\/code> telling you it happened. On an oversubscribed VPS that is the difference between a degraded service and a dead one. Pair it with correct application-level sizing \u2014 see our <a href=\"https:\/\/virtualserversvps.com\/blog\/tuning-php-fpm-process-pools-by-workload\">PHP-FPM pool sizing guide<\/a> \u2014 and you have a system that fails locally instead of globally.<\/p>\n<!-- \/wp:paragraph -->\n<!-- \/wp:list-item --><li><strong>Setting only <code>memory.max<\/code>.<\/strong> Without <code>memory.high<\/code> you lose the graceful throttle and go straight to a kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Forgetting swap.<\/strong> <code>memory.swap.max=0<\/code> on a database is often correct, but on a stateless web worker it turns manageable pressure into a hard kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Ignoring nested cgroups.<\/strong> A limit on a parent slice and a child unit both apply; the more restrictive one wins. Check the full path with <code>systemd-cgls<\/code>.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Trusting old guides.<\/strong> Anything referencing <code>memory.limit_in_bytes<\/code>, <code>tasks<\/code> or <code>cgroup.clone_children<\/code> is cgroups v1 and will not work on a modern VPS.<\/li><!-- \/wp:list-item --><\/ul>\n<!-- \/wp:list -->\n\n<!-- wp:heading {\"level\":2} -->\n<h2>Where this fits<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>cgroup memory limits do not make a workload fit in less RAM. They make the failure mode predictable: instead of a global OOM that takes out MariaDB because it happened to be the largest RSS, you get a scoped kill of the process that actually misbehaved, and a counter in <code>memory.events<\/code> telling you it happened. On an oversubscribed VPS that is the difference between a degraded service and a dead one. Pair it with correct application-level sizing \u2014 see our <a href=\"https:\/\/virtualserversvps.com\/blog\/tuning-php-fpm-process-pools-by-workload\">PHP-FPM pool sizing guide<\/a> \u2014 and you have a system that fails locally instead of globally.<\/p>\n<!-- \/wp:paragraph -->\n<!-- \/wp:list-item --><!-- wp:list-item --><li><strong>Setting only <code>memory.max<\/code>.<\/strong> Without <code>memory.high<\/code> you lose the graceful throttle and go straight to a kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Forgetting swap.<\/strong> <code>memory.swap.max=0<\/code> on a database is often correct, but on a stateless web worker it turns manageable pressure into a hard kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Ignoring nested cgroups.<\/strong> A limit on a parent slice and a child unit both apply; the more restrictive one wins. Check the full path with <code>systemd-cgls<\/code>.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Trusting old guides.<\/strong> Anything referencing <code>memory.limit_in_bytes<\/code>, <code>tasks<\/code> or <code>cgroup.clone_children<\/code> is cgroups v1 and will not work on a modern VPS.<\/li><!-- \/wp:list-item --><\/ul>\n<!-- \/wp:list -->\n\n<!-- wp:heading {\"level\":2} -->\n<h2>Where this fits<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>cgroup memory limits do not make a workload fit in less RAM. They make the failure mode predictable: instead of a global OOM that takes out MariaDB because it happened to be the largest RSS, you get a scoped kill of the process that actually misbehaved, and a counter in <code>memory.events<\/code> telling you it happened. On an oversubscribed VPS that is the difference between a degraded service and a dead one. Pair it with correct application-level sizing \u2014 see our <a href=\"https:\/\/virtualserversvps.com\/blog\/tuning-php-fpm-process-pools-by-workload\">PHP-FPM pool sizing guide<\/a> \u2014 and you have a system that fails locally instead of globally.<\/p>\n<!-- \/wp:paragraph -->\n<!-- \/wp:list-item -->\n\n<!-- wp:list -->\n<ul><!-- wp:list-item --><li><strong>Assuming <code>memory.max<\/code> caps RSS.<\/strong> It caps the cgroup&#8217;s total memory including page cache charged to it. A process reading a 300&nbsp;MB file may hit its limit with a small RSS.<\/li><!-- \/wp:list-item --><!-- wp:list-item --><li><strong>Setting only <code>memory.max<\/code>.<\/strong> Without <code>memory.high<\/code> you lose the graceful throttle and go straight to a kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Forgetting swap.<\/strong> <code>memory.swap.max=0<\/code> on a database is often correct, but on a stateless web worker it turns manageable pressure into a hard kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Ignoring nested cgroups.<\/strong> A limit on a parent slice and a child unit both apply; the more restrictive one wins. Check the full path with <code>systemd-cgls<\/code>.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Trusting old guides.<\/strong> Anything referencing <code>memory.limit_in_bytes<\/code>, <code>tasks<\/code> or <code>cgroup.clone_children<\/code> is cgroups v1 and will not work on a modern VPS.<\/li><!-- \/wp:list-item --><\/ul>\n<!-- \/wp:list -->\n\n<!-- wp:heading {\"level\":2} -->\n<h2>Where this fits<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>cgroup memory limits do not make a workload fit in less RAM. They make the failure mode predictable: instead of a global OOM that takes out MariaDB because it happened to be the largest RSS, you get a scoped kill of the process that actually misbehaved, and a counter in <code>memory.events<\/code> telling you it happened. On an oversubscribed VPS that is the difference between a degraded service and a dead one. Pair it with correct application-level sizing \u2014 see our <a href=\"https:\/\/virtualserversvps.com\/blog\/tuning-php-fpm-process-pools-by-workload\">PHP-FPM pool sizing guide<\/a> \u2014 and you have a system that fails locally instead of globally.<\/p>\n<!-- \/wp:paragraph -->\n<!-- \/wp:list --><!-- \/wp:post-content --><!-- \/wp:list-item --><li><strong>Forgetting swap.<\/strong> <code>memory.swap.max=0<\/code> on a database is often correct, but on a stateless web worker it turns manageable pressure into a hard kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Ignoring nested cgroups.<\/strong> A limit on a parent slice and a child unit both apply; the more restrictive one wins. Check the full path with <code>systemd-cgls<\/code>.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Trusting old guides.<\/strong> Anything referencing <code>memory.limit_in_bytes<\/code>, <code>tasks<\/code> or <code>cgroup.clone_children<\/code> is cgroups v1 and will not work on a modern VPS.<\/li><!-- \/wp:list-item --><\/ul>\n<!-- \/wp:list-item -->\n\n<!-- wp:heading {\"level\":2} -->\n<h2>Where this fits<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>cgroup memory limits do not make a workload fit in less RAM. They make the failure mode predictable: instead of a global OOM that takes out MariaDB because it happened to be the largest RSS, you get a scoped kill of the process that actually misbehaved, and a counter in <code>memory.events<\/code> telling you it happened. On an oversubscribed VPS that is the difference between a degraded service and a dead one. Pair it with correct application-level sizing \u2014 see our <a href=\"https:\/\/virtualserversvps.com\/blog\/tuning-php-fpm-process-pools-by-workload\">PHP-FPM pool sizing guide<\/a> \u2014 and you have a system that fails locally instead of globally.<\/p>\n<!-- \/wp:paragraph -->\n<!-- \/wp:list-item --><!-- wp:list-item --><li><strong>Forgetting swap.<\/strong> <code>memory.swap.max=0<\/code> on a database is often correct, but on a stateless web worker it turns manageable pressure into a hard kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Ignoring nested cgroups.<\/strong> A limit on a parent slice and a child unit both apply; the more restrictive one wins. Check the full path with <code>systemd-cgls<\/code>.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Trusting old guides.<\/strong> Anything referencing <code>memory.limit_in_bytes<\/code>, <code>tasks<\/code> or <code>cgroup.clone_children<\/code> is cgroups v1 and will not work on a modern VPS.<\/li><!-- \/wp:list-item --><\/ul>\n<!-- \/wp:list -->\n\n<!-- wp:heading {\"level\":2} -->\n<h2>Where this fits<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>cgroup memory limits do not make a workload fit in less RAM. They make the failure mode predictable: instead of a global OOM that takes out MariaDB because it happened to be the largest RSS, you get a scoped kill of the process that actually misbehaved, and a counter in <code>memory.events<\/code> telling you it happened. On an oversubscribed VPS that is the difference between a degraded service and a dead one. Pair it with correct application-level sizing \u2014 see our <a href=\"https:\/\/virtualserversvps.com\/blog\/tuning-php-fpm-process-pools-by-workload\">PHP-FPM pool sizing guide<\/a> \u2014 and you have a system that fails locally instead of globally.<\/p>\n<!-- \/wp:paragraph -->\n<!-- \/wp:list-item --><li><strong>Setting only <code>memory.max<\/code>.<\/strong> Without <code>memory.high<\/code> you lose the graceful throttle and go straight to a kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Forgetting swap.<\/strong> <code>memory.swap.max=0<\/code> on a database is often correct, but on a stateless web worker it turns manageable pressure into a hard kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Ignoring nested cgroups.<\/strong> A limit on a parent slice and a child unit both apply; the more restrictive one wins. Check the full path with <code>systemd-cgls<\/code>.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Trusting old guides.<\/strong> Anything referencing <code>memory.limit_in_bytes<\/code>, <code>tasks<\/code> or <code>cgroup.clone_children<\/code> is cgroups v1 and will not work on a modern VPS.<\/li><!-- \/wp:list-item --><\/ul>\n<!-- \/wp:list -->\n\n<!-- wp:heading {\"level\":2} -->\n<h2>Where this fits<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>cgroup memory limits do not make a workload fit in less RAM. They make the failure mode predictable: instead of a global OOM that takes out MariaDB because it happened to be the largest RSS, you get a scoped kill of the process that actually misbehaved, and a counter in <code>memory.events<\/code> telling you it happened. On an oversubscribed VPS that is the difference between a degraded service and a dead one. Pair it with correct application-level sizing \u2014 see our <a href=\"https:\/\/virtualserversvps.com\/blog\/tuning-php-fpm-process-pools-by-workload\">PHP-FPM pool sizing guide<\/a> \u2014 and you have a system that fails locally instead of globally.<\/p>\n<!-- \/wp:paragraph -->\n<!-- \/wp:list-item --><!-- wp:list-item --><li><strong>Setting only <code>memory.max<\/code>.<\/strong> Without <code>memory.high<\/code> you lose the graceful throttle and go straight to a kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Forgetting swap.<\/strong> <code>memory.swap.max=0<\/code> on a database is often correct, but on a stateless web worker it turns manageable pressure into a hard kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Ignoring nested cgroups.<\/strong> A limit on a parent slice and a child unit both apply; the more restrictive one wins. Check the full path with <code>systemd-cgls<\/code>.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Trusting old guides.<\/strong> Anything referencing <code>memory.limit_in_bytes<\/code>, <code>tasks<\/code> or <code>cgroup.clone_children<\/code> is cgroups v1 and will not work on a modern VPS.<\/li><!-- \/wp:list-item --><\/ul>\n<!-- \/wp:list -->\n\n<!-- wp:heading {\"level\":2} -->\n<h2>Where this fits<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>cgroup memory limits do not make a workload fit in less RAM. They make the failure mode predictable: instead of a global OOM that takes out MariaDB because it happened to be the largest RSS, you get a scoped kill of the process that actually misbehaved, and a counter in <code>memory.events<\/code> telling you it happened. On an oversubscribed VPS that is the difference between a degraded service and a dead one. Pair it with correct application-level sizing \u2014 see our <a href=\"https:\/\/virtualserversvps.com\/blog\/tuning-php-fpm-process-pools-by-workload\">PHP-FPM pool sizing guide<\/a> \u2014 and you have a system that fails locally instead of globally.<\/p>\n<!-- \/wp:paragraph -->\n<!-- \/wp:list-item -->\n\n<!-- wp:list -->\n<ul><!-- wp:list-item --><li><strong>Assuming <code>memory.max<\/code> caps RSS.<\/strong> It caps the cgroup&#8217;s total memory including page cache charged to it. A process reading a 300&nbsp;MB file may hit its limit with a small RSS.<\/li><!-- \/wp:list-item --><!-- wp:list-item --><li><strong>Setting only <code>memory.max<\/code>.<\/strong> Without <code>memory.high<\/code> you lose the graceful throttle and go straight to a kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Forgetting swap.<\/strong> <code>memory.swap.max=0<\/code> on a database is often correct, but on a stateless web worker it turns manageable pressure into a hard kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Ignoring nested cgroups.<\/strong> A limit on a parent slice and a child unit both apply; the more restrictive one wins. Check the full path with <code>systemd-cgls<\/code>.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Trusting old guides.<\/strong> Anything referencing <code>memory.limit_in_bytes<\/code>, <code>tasks<\/code> or <code>cgroup.clone_children<\/code> is cgroups v1 and will not work on a modern VPS.<\/li><!-- \/wp:list-item --><\/ul>\n<!-- \/wp:list -->\n\n<!-- wp:heading {\"level\":2} -->\n<h2>Where this fits<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>cgroup memory limits do not make a workload fit in less RAM. They make the failure mode predictable: instead of a global OOM that takes out MariaDB because it happened to be the largest RSS, you get a scoped kill of the process that actually misbehaved, and a counter in <code>memory.events<\/code> telling you it happened. On an oversubscribed VPS that is the difference between a degraded service and a dead one. Pair it with correct application-level sizing \u2014 see our <a href=\"https:\/\/virtualserversvps.com\/blog\/tuning-php-fpm-process-pools-by-workload\">PHP-FPM pool sizing guide<\/a> \u2014 and you have a system that fails locally instead of globally.<\/p>\n<!-- \/wp:paragraph -->\n<!-- \/wp:list --><!-- \/wp:post-content --><!-- \/wp:list-item --><!-- wp:post-content --><!-- wp:paragraph -->\n<p>On a VPS with 1&nbsp;GB of RAM you have two blunt instruments for memory pressure: point your applications at sane limits and hope, or let the kernel OOM killer decide. Neither is satisfying, and the OOM killer&#8217;s choice is frequently not the process you would have picked. cgroups v2 gives you a third option \u2014 a hard, enforceable ceiling on any process tree, with the OOM event delivered to that cgroup instead of the whole system. This article covers setting that up from scratch on a Debian 12 \/ Ubuntu 24.04 VPS, and the differences from cgroups v1 that catch people out.<\/p>\n<!-- \/wp:paragraph -->\n\n<!-- wp:paragraph -->\n<p>Before capping anything, it helps to know what your workload&#8217;s real working set is. Our <a href=\"https:\/\/virtualserversvps.com\/blog\/load-average-vs-cpu-saturation-vps-monitoring\">guide to reading load average versus actual CPU saturation<\/a> explains why memory pressure and load average tell different stories on a shared VPS.<\/p>\n<!-- \/wp:paragraph -->\n\n<!-- wp:heading {\"level\":2} -->\n<h2>Confirm you are on cgroups v2<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p><code>stat -fc %T \/sys\/fs\/cgroup<\/code><\/p>\n<!-- \/wp:paragraph -->\n\n<!-- wp:paragraph -->\n<p><code>cgroup2fs<\/code> means unified v2. <code>tmpfs<\/code> means v1 and this article&#8217;s paths do not apply. Check the controller list:<\/p>\n<!-- \/wp:paragraph -->\n\n<!-- wp:paragraph -->\n<p><code>cat \/sys\/fs\/cgroup\/cgroup.controllers<\/code><\/p>\n<!-- \/wp:paragraph -->\n\n<!-- wp:paragraph -->\n<p>You should see at least <code>cpu io memory pids<\/code>. On Ubuntu the <code>memory<\/code> controller is available at the root but must be enabled in child cgroups by writing to <code>cgroup.subtree_control<\/code>. Debian 12 with a recent kernel behaves the same.<\/p>\n<!-- \/wp:paragraph -->\n\n<!-- wp:heading {\"level\":2} -->\n<h2>The three memory knobs that matter<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:table -->\n<figure class=\"wp-block-table\"><table><thead><tr><th>File<\/th><th>Type<\/th><th>Behaviour<\/th><th>Typical use<\/th><\/tr><\/thead><tbody><tr><td><code>memory.max<\/code><\/td><td>Hard limit<\/td><td>Allocation triggers reclaim; if reclaim fails, cgroup OOM killer fires<\/td><td>Hard ceiling for an app that must not starve the host<\/td><\/tr><tr><td><code>memory.high<\/code><\/td><td>Soft limit<\/td><td>Throttles allocation and reclaims aggressively, no OOM<\/td><td>Preferred: gives the app a chance to shed load<\/td><\/tr><tr><td><code>memory.min<\/code><\/td><td>Protection<\/td><td>Pages here are never reclaimed under pressure<\/td><td>Protect a critical daemon in a shared cgroup tree<\/td><\/tr><tr><td><code>memory.swap.max<\/code><\/td><td>Hard limit<\/td><td>Caps swap the cgroup may use<\/td><td>Stop one service from swapping the whole box<\/td><\/tr><tr><td><code>memory.low<\/code><\/td><td>Protection (soft)<\/td><td>Reclaim avoids this range where possible<\/td><td>Softer alternative to <code>memory.min<\/code><\/td><\/tr><tr><td><code>memory.oom.group<\/code><\/td><td>Toggle<\/td><td>Kill the entire cgroup instead of one process<\/td><td>Multi-process apps that corrupt state if half-killed<\/td><\/tr><\/tbody><\/table><\/figure>\n<!-- \/wp:table -->\n\n<!-- wp:paragraph -->\n<p>The practical advice: set <code>memory.high<\/code> at roughly 90% of your intended ceiling and <code>memory.max<\/code> at 100%. The high watermark throttles and reclaims gradually, which most applications survive; the max watermark is the backstop that prevents a runaway from consuming the host.<\/p>\n<!-- \/wp:paragraph -->\n\n<!-- wp:heading {\"level\":2} -->\n<h2>Cap a service manually with systemd<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>systemd already creates a cgroup per unit and exposes the controls as directives. This is the least error-prone route and needs no manual filesystem work. Create a drop-in:<\/p>\n<!-- \/wp:paragraph -->\n\n<!-- wp:paragraph -->\n<p><code>systemctl edit php8.3-fpm.service<\/code><\/p>\n<!-- \/wp:paragraph -->\n\n<!-- wp:paragraph -->\n<p><code>[Service]<br>MemoryHigh=460M<br>MemoryMax=512M<br>MemorySwapMax=128M<br>MemoryAccounting=yes<br>TasksMax=256<\/code><\/p>\n<!-- \/wp:paragraph -->\n\n<!-- wp:paragraph -->\n<p><code>systemctl daemon-reload &amp;&amp; systemctl restart php8.3-fpm.service<\/code><\/p>\n<!-- \/wp:paragraph -->\n\n<!-- wp:paragraph -->\n<p>Verify:<\/p>\n<!-- \/wp:paragraph -->\n\n<!-- wp:paragraph -->\n<p><code>systemctl show php8.3-fpm.service -p MemoryCurrent -p MemoryMax -p MemoryHigh<br>cat \/sys\/fs\/cgroup\/system.slice\/php8.3-fpm.service\/memory.current<\/code><\/p>\n<!-- \/wp:paragraph -->\n\n<!-- wp:paragraph -->\n<p>Note that <code>MemoryCurrent<\/code> is in bytes and reports <code>[not set]<\/code> if <code>MemoryAccounting=<\/code> is off. Always set accounting explicitly.<\/p>\n<!-- \/wp:paragraph -->\n\n<!-- wp:heading {\"level\":2} -->\n<h2>Cap an arbitrary process tree with systemd-run<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>For something not managed by a unit \u2014 a batch job, a scraper, a build \u2014 wrap it:<\/p>\n<!-- \/wp:paragraph -->\n\n<!-- wp:paragraph -->\n<p><code>systemd-run --scope --unit=batchjob --property=MemoryMax=384M --property=MemorySwapMax=0 \\<br>&nbsp;&nbsp;\/usr\/local\/bin\/nightly-indexer.sh<\/code><\/p>\n<!-- \/wp:paragraph -->\n\n<!-- wp:paragraph -->\n<p><code>--scope<\/code> runs the process in the current session but still inside a cgroup, so the cap applies and it appears in <code>systemd-cgtop<\/code>. If the job exceeds 384&nbsp;MB it is killed by its own cgroup OOM, and the rest of the VPS is unaffected. This is by far the simplest way to stop a one-off job from taking down your database.<\/p>\n<!-- \/wp:paragraph -->\n\n<!-- wp:heading {\"level\":2} -->\n<h2>Watching pressure and OOM events<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>Every cgroup exposes its own pressure stall information. For a cgroup at <code>\/sys\/fs\/cgroup\/system.slice\/php8.3-fpm.service\/<\/code>:<\/p>\n<!-- \/wp:paragraph -->\n\n<!-- wp:paragraph -->\n<p><code>cat memory.pressure<br>cat memory.events<\/code><\/p>\n<!-- \/wp:paragraph -->\n\n<!-- wp:paragraph -->\n<p><code>memory.events<\/code> counts <code>high<\/code>, <code>max<\/code>, <code>oom<\/code> and <code>oom_kill<\/code> occurrences since boot. A rising <code>high<\/code> count with zero <code>oom_kill<\/code> is exactly what you want: the throttle is working. Non-zero <code>oom_kill<\/code> means the ceiling is genuinely too low for the workload, not that tuning is needed.<\/p>\n<!-- \/wp:paragraph -->\n\n<!-- wp:paragraph -->\n<p><code>systemd-cgtop -m --order=memory<\/code> gives a live ranked view, which is usually faster than reading the files by hand when you suspect a leak.<\/p>\n<!-- \/wp:paragraph -->\n\n<!-- wp:heading {\"level\":2} -->\n<h2>A worked oversubscription plan for a 1 GB VPS<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>Assume 1&nbsp;GB total, ~120&nbsp;MB consumed by the kernel and base system. That leaves roughly 880&nbsp;MB to allocate. Deliberately oversubscribe the soft limits (they can borrow when idle) but keep the hard limits within the physical budget.<\/p>\n<!-- \/wp:paragraph -->\n\n<!-- wp:table -->\n<figure class=\"wp-block-table\"><table><thead><tr><th>Service<\/th><th>memory.high<\/th><th>memory.max<\/th><th>swap.max<\/th><th>Rationale<\/th><\/tr><\/thead><tbody><tr><td>MariaDB<\/td><td>320M<\/td><td>380M<\/td><td>64M<\/td><td>Buffer pool is fixed; small headroom for connections<\/td><\/tr><tr><td>PHP-FPM<\/td><td>200M<\/td><td>240M<\/td><td>64M<\/td><td>Scales with children; throttle before OOM<\/td><\/tr><tr><td>Nginx<\/td><td>24M<\/td><td>48M<\/td><td>0<\/td><td>Tiny, should never swap<\/td><\/tr><tr><td>Redis<\/td><td>64M<\/td><td>96M<\/td><td>0<\/td><td>Set <code>maxmemory<\/code> in Redis too<\/td><\/tr><tr><td>System \/ sshd \/ cron<\/td><td>96M<\/td><td>128M<\/td><td>32M<\/td><td>Protected with <code>memory.min<\/code><\/td><\/tr><tr><td><strong>Total hard max<\/strong><\/td><td>\u2014<\/td><td><strong>892M<\/strong><\/td><td>160M<\/td><td>Fits in RAM, swap is a cushion only<\/td><\/tr><\/tbody><\/table><\/figure>\n<!-- \/wp:table -->\n\n<!-- wp:paragraph -->\n<p>Protect the system slice so SSH stays reachable even when MariaDB is reclaiming hard:<\/p>\n<!-- \/wp:paragraph -->\n\n<!-- wp:paragraph -->\n<p><code>echo 67108864 &gt; \/sys\/fs\/cgroup\/system.slice\/memory.min<\/code><\/p>\n<!-- \/wp:paragraph -->\n\n<!-- wp:paragraph -->\n<p>Persist that with a drop-in on <code>system.slice<\/code> or, more practically, accept the systemd defaults and instead make sure your application limits are the ones being hit first. Setting a <code>memory.min<\/code> on the system slice is a blunt safety net and worth doing once you have been OOM-killed out of an SSH session during an incident.<\/p>\n<!-- \/wp:paragraph -->\n\n<!-- wp:heading {\"level\":2} -->\n<h2>Common mistakes<\/h2>\n<!-- \/wp:heading --><!-- wp:list-item --><li><strong>Ignoring nested cgroups.<\/strong> A limit on a parent slice and a child unit both apply; the more restrictive one wins. Check the full path with <code>systemd-cgls<\/code>.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Trusting old guides.<\/strong> Anything referencing <code>memory.limit_in_bytes<\/code>, <code>tasks<\/code> or <code>cgroup.clone_children<\/code> is cgroups v1 and will not work on a modern VPS.<\/li><!-- \/wp:list-item --><\/ul>\n<!-- \/wp:list-item -->\n\n<!-- wp:heading {\"level\":2} -->\n<h2>Where this fits<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>cgroup memory limits do not make a workload fit in less RAM. They make the failure mode predictable: instead of a global OOM that takes out MariaDB because it happened to be the largest RSS, you get a scoped kill of the process that actually misbehaved, and a counter in <code>memory.events<\/code> telling you it happened. On an oversubscribed VPS that is the difference between a degraded service and a dead one. Pair it with correct application-level sizing \u2014 see our <a href=\"https:\/\/virtualserversvps.com\/blog\/tuning-php-fpm-process-pools-by-workload\">PHP-FPM pool sizing guide<\/a> \u2014 and you have a system that fails locally instead of globally.<\/p>\n<!-- \/wp:paragraph -->\n<!-- \/wp:list-item --><li><strong>Forgetting swap.<\/strong> <code>memory.swap.max=0<\/code> on a database is often correct, but on a stateless web worker it turns manageable pressure into a hard kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Ignoring nested cgroups.<\/strong> A limit on a parent slice and a child unit both apply; the more restrictive one wins. Check the full path with <code>systemd-cgls<\/code>.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Trusting old guides.<\/strong> Anything referencing <code>memory.limit_in_bytes<\/code>, <code>tasks<\/code> or <code>cgroup.clone_children<\/code> is cgroups v1 and will not work on a modern VPS.<\/li><!-- \/wp:list-item --><\/ul>\n<!-- \/wp:list -->\n\n<!-- wp:heading {\"level\":2} -->\n<h2>Where this fits<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>cgroup memory limits do not make a workload fit in less RAM. They make the failure mode predictable: instead of a global OOM that takes out MariaDB because it happened to be the largest RSS, you get a scoped kill of the process that actually misbehaved, and a counter in <code>memory.events<\/code> telling you it happened. On an oversubscribed VPS that is the difference between a degraded service and a dead one. Pair it with correct application-level sizing \u2014 see our <a href=\"https:\/\/virtualserversvps.com\/blog\/tuning-php-fpm-process-pools-by-workload\">PHP-FPM pool sizing guide<\/a> \u2014 and you have a system that fails locally instead of globally.<\/p>\n<!-- \/wp:paragraph -->\n<!-- \/wp:list-item --><!-- wp:list-item --><li><strong>Forgetting swap.<\/strong> <code>memory.swap.max=0<\/code> on a database is often correct, but on a stateless web worker it turns manageable pressure into a hard kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Ignoring nested cgroups.<\/strong> A limit on a parent slice and a child unit both apply; the more restrictive one wins. Check the full path with <code>systemd-cgls<\/code>.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Trusting old guides.<\/strong> Anything referencing <code>memory.limit_in_bytes<\/code>, <code>tasks<\/code> or <code>cgroup.clone_children<\/code> is cgroups v1 and will not work on a modern VPS.<\/li><!-- \/wp:list-item --><\/ul>\n<!-- \/wp:list -->\n\n<!-- wp:heading {\"level\":2} -->\n<h2>Where this fits<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>cgroup memory limits do not make a workload fit in less RAM. They make the failure mode predictable: instead of a global OOM that takes out MariaDB because it happened to be the largest RSS, you get a scoped kill of the process that actually misbehaved, and a counter in <code>memory.events<\/code> telling you it happened. On an oversubscribed VPS that is the difference between a degraded service and a dead one. Pair it with correct application-level sizing \u2014 see our <a href=\"https:\/\/virtualserversvps.com\/blog\/tuning-php-fpm-process-pools-by-workload\">PHP-FPM pool sizing guide<\/a> \u2014 and you have a system that fails locally instead of globally.<\/p>\n<!-- \/wp:paragraph -->\n<!-- \/wp:list-item --><li><strong>Setting only <code>memory.max<\/code>.<\/strong> Without <code>memory.high<\/code> you lose the graceful throttle and go straight to a kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Forgetting swap.<\/strong> <code>memory.swap.max=0<\/code> on a database is often correct, but on a stateless web worker it turns manageable pressure into a hard kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Ignoring nested cgroups.<\/strong> A limit on a parent slice and a child unit both apply; the more restrictive one wins. Check the full path with <code>systemd-cgls<\/code>.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Trusting old guides.<\/strong> Anything referencing <code>memory.limit_in_bytes<\/code>, <code>tasks<\/code> or <code>cgroup.clone_children<\/code> is cgroups v1 and will not work on a modern VPS.<\/li><!-- \/wp:list-item --><\/ul>\n<!-- \/wp:list -->\n\n<!-- wp:heading {\"level\":2} -->\n<h2>Where this fits<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>cgroup memory limits do not make a workload fit in less RAM. They make the failure mode predictable: instead of a global OOM that takes out MariaDB because it happened to be the largest RSS, you get a scoped kill of the process that actually misbehaved, and a counter in <code>memory.events<\/code> telling you it happened. On an oversubscribed VPS that is the difference between a degraded service and a dead one. Pair it with correct application-level sizing \u2014 see our <a href=\"https:\/\/virtualserversvps.com\/blog\/tuning-php-fpm-process-pools-by-workload\">PHP-FPM pool sizing guide<\/a> \u2014 and you have a system that fails locally instead of globally.<\/p>\n<!-- \/wp:paragraph -->\n<!-- \/wp:list-item --><!-- wp:list-item --><li><strong>Setting only <code>memory.max<\/code>.<\/strong> Without <code>memory.high<\/code> you lose the graceful throttle and go straight to a kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Forgetting swap.<\/strong> <code>memory.swap.max=0<\/code> on a database is often correct, but on a stateless web worker it turns manageable pressure into a hard kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Ignoring nested cgroups.<\/strong> A limit on a parent slice and a child unit both apply; the more restrictive one wins. Check the full path with <code>systemd-cgls<\/code>.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Trusting old guides.<\/strong> Anything referencing <code>memory.limit_in_bytes<\/code>, <code>tasks<\/code> or <code>cgroup.clone_children<\/code> is cgroups v1 and will not work on a modern VPS.<\/li><!-- \/wp:list-item --><\/ul>\n<!-- \/wp:list -->\n\n<!-- wp:heading {\"level\":2} -->\n<h2>Where this fits<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>cgroup memory limits do not make a workload fit in less RAM. They make the failure mode predictable: instead of a global OOM that takes out MariaDB because it happened to be the largest RSS, you get a scoped kill of the process that actually misbehaved, and a counter in <code>memory.events<\/code> telling you it happened. On an oversubscribed VPS that is the difference between a degraded service and a dead one. Pair it with correct application-level sizing \u2014 see our <a href=\"https:\/\/virtualserversvps.com\/blog\/tuning-php-fpm-process-pools-by-workload\">PHP-FPM pool sizing guide<\/a> \u2014 and you have a system that fails locally instead of globally.<\/p>\n<!-- \/wp:paragraph -->\n<!-- \/wp:list-item -->\n\n<!-- wp:list -->\n<ul><!-- wp:list-item --><li><strong>Assuming <code>memory.max<\/code> caps RSS.<\/strong> It caps the cgroup&#8217;s total memory including page cache charged to it. A process reading a 300&nbsp;MB file may hit its limit with a small RSS.<\/li><!-- \/wp:list-item --><!-- wp:list-item --><li><strong>Setting only <code>memory.max<\/code>.<\/strong> Without <code>memory.high<\/code> you lose the graceful throttle and go straight to a kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Forgetting swap.<\/strong> <code>memory.swap.max=0<\/code> on a database is often correct, but on a stateless web worker it turns manageable pressure into a hard kill.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Ignoring nested cgroups.<\/strong> A limit on a parent slice and a child unit both apply; the more restrictive one wins. Check the full path with <code>systemd-cgls<\/code>.<\/li><!-- wp:list-item --><!-- wp:list-item --><li><strong>Trusting old guides.<\/strong> Anything referencing <code>memory.limit_in_bytes<\/code>, <code>tasks<\/code> or <code>cgroup.clone_children<\/code> is cgroups v1 and will not work on a modern VPS.<\/li><!-- \/wp:list-item --><\/ul>\n<!-- \/wp:list -->\n\n<!-- wp:heading {\"level\":2} -->\n<h2>Where this fits<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>cgroup memory limits do not make a workload fit in less RAM. They make the failure mode predictable: instead of a global OOM that takes out MariaDB because it happened to be the largest RSS, you get a scoped kill of the process that actually misbehaved, and a counter in <code>memory.events<\/code> telling you it happened. On an oversubscribed VPS that is the difference between a degraded service and a dead one. Pair it with correct application-level sizing \u2014 see our <a href=\"https:\/\/virtualserversvps.com\/blog\/tuning-php-fpm-process-pools-by-workload\">PHP-FPM pool sizing guide<\/a> \u2014 and you have a system that fails locally instead of globally.<\/p>\n<!-- \/wp:paragraph -->\n<!-- \/wp:list --><!-- \/wp:post-content -->","protected":false},"excerpt":{"rendered":"<p>Hard and soft memory ceilings per service with cgroups v2 on Debian\/Ubuntu VPS, plus a worked allocation plan for a 1 GB instance.<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"iawp_total_views":0,"footnotes":""},"categories":[2],"tags":[],"class_list":["post-1188","post","type-post","status-publish","format-standard","hentry","category-provider-reviews-comparisons"],"yoast_head":"<!-- This site is optimized with the Yoast SEO Premium plugin v26.1 (Yoast SEO v26.1) - https:\/\/yoast.com\/wordpress\/plugins\/seo\/ -->\n<title>Capping Memory with cgroups v2 on an Oversubscribed VPS - Virtual Servers VPS Blog<\/title>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/virtualserversvps.com\/blog\/cgroups-v2-cap-memory-oversubscribed-vps\/\" \/>\n<meta property=\"og:locale\" content=\"en_US\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"Capping Memory with cgroups v2 on an Oversubscribed VPS\" \/>\n<meta property=\"og:description\" content=\"Capping Memory with cgroups v2 on an Oversubscribed VPS\" \/>\n<meta property=\"og:url\" content=\"https:\/\/virtualserversvps.com\/blog\/cgroups-v2-cap-memory-oversubscribed-vps\/\" \/>\n<meta property=\"og:site_name\" content=\"Virtual Servers VPS Blog\" \/>\n<meta property=\"article:published_time\" content=\"2026-09-20T22:36:45+00:00\" \/>\n<meta name=\"author\" content=\"Virtual-Servers-Vps-Editor\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"Written by\" \/>\n\t<meta name=\"twitter:data1\" content=\"Virtual-Servers-Vps-Editor\" \/>\n\t<meta name=\"twitter:label2\" content=\"Est. reading time\" \/>\n\t<meta name=\"twitter:data2\" content=\"10 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"WebPage\",\"@id\":\"https:\/\/virtualserversvps.com\/blog\/cgroups-v2-cap-memory-oversubscribed-vps\/\",\"url\":\"https:\/\/virtualserversvps.com\/blog\/cgroups-v2-cap-memory-oversubscribed-vps\/\",\"name\":\"Capping Memory with cgroups v2 on an Oversubscribed VPS - Virtual Servers VPS Blog\",\"isPartOf\":{\"@id\":\"https:\/\/virtualserversvps.com\/blog\/#website\"},\"datePublished\":\"2026-09-20T22:36:45+00:00\",\"author\":{\"@id\":\"https:\/\/virtualserversvps.com\/blog\/#\/schema\/person\/82a299a8284a66ff49f97c74684724a0\"},\"breadcrumb\":{\"@id\":\"https:\/\/virtualserversvps.com\/blog\/cgroups-v2-cap-memory-oversubscribed-vps\/#breadcrumb\"},\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/virtualserversvps.com\/blog\/cgroups-v2-cap-memory-oversubscribed-vps\/\"]}]},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/virtualserversvps.com\/blog\/cgroups-v2-cap-memory-oversubscribed-vps\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\/\/virtualserversvps.com\/blog\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"Capping Memory with cgroups v2 on an Oversubscribed VPS\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/virtualserversvps.com\/blog\/#website\",\"url\":\"https:\/\/virtualserversvps.com\/blog\/\",\"name\":\"Virtual Servers VPS Blog\",\"description\":\"\",\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/virtualserversvps.com\/blog\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"en-US\"},{\"@type\":\"Person\",\"@id\":\"https:\/\/virtualserversvps.com\/blog\/#\/schema\/person\/82a299a8284a66ff49f97c74684724a0\",\"name\":\"Virtual-Servers-Vps-Editor\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\/\/virtualserversvps.com\/blog\/#\/schema\/person\/image\/\",\"url\":\"https:\/\/secure.gravatar.com\/avatar\/d820b15f1cd028e97610d9adf536df7be5cb6423869967037d468d5355fa003f?s=96&d=mm&r=g\",\"contentUrl\":\"https:\/\/secure.gravatar.com\/avatar\/d820b15f1cd028e97610d9adf536df7be5cb6423869967037d468d5355fa003f?s=96&d=mm&r=g\",\"caption\":\"Virtual-Servers-Vps-Editor\"},\"sameAs\":[\"https:\/\/virtualserversvps.com\/blog\"],\"url\":\"https:\/\/virtualserversvps.com\/blog\/author\/virtualserversvps\/\"}]}<\/script>\n<!-- \/ Yoast SEO Premium plugin. -->","yoast_head_json":{"title":"Capping Memory with cgroups v2 on an Oversubscribed VPS - Virtual Servers VPS Blog","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/virtualserversvps.com\/blog\/cgroups-v2-cap-memory-oversubscribed-vps\/","og_locale":"en_US","og_type":"article","og_title":"Capping Memory with cgroups v2 on an Oversubscribed VPS","og_description":"Capping Memory with cgroups v2 on an Oversubscribed VPS","og_url":"https:\/\/virtualserversvps.com\/blog\/cgroups-v2-cap-memory-oversubscribed-vps\/","og_site_name":"Virtual Servers VPS Blog","article_published_time":"2026-09-20T22:36:45+00:00","author":"Virtual-Servers-Vps-Editor","twitter_card":"summary_large_image","twitter_misc":{"Written by":"Virtual-Servers-Vps-Editor","Est. reading time":"10 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"WebPage","@id":"https:\/\/virtualserversvps.com\/blog\/cgroups-v2-cap-memory-oversubscribed-vps\/","url":"https:\/\/virtualserversvps.com\/blog\/cgroups-v2-cap-memory-oversubscribed-vps\/","name":"Capping Memory with cgroups v2 on an Oversubscribed VPS - Virtual Servers VPS Blog","isPartOf":{"@id":"https:\/\/virtualserversvps.com\/blog\/#website"},"datePublished":"2026-09-20T22:36:45+00:00","author":{"@id":"https:\/\/virtualserversvps.com\/blog\/#\/schema\/person\/82a299a8284a66ff49f97c74684724a0"},"breadcrumb":{"@id":"https:\/\/virtualserversvps.com\/blog\/cgroups-v2-cap-memory-oversubscribed-vps\/#breadcrumb"},"inLanguage":"en-US","potentialAction":[{"@type":"ReadAction","target":["https:\/\/virtualserversvps.com\/blog\/cgroups-v2-cap-memory-oversubscribed-vps\/"]}]},{"@type":"BreadcrumbList","@id":"https:\/\/virtualserversvps.com\/blog\/cgroups-v2-cap-memory-oversubscribed-vps\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/virtualserversvps.com\/blog\/"},{"@type":"ListItem","position":2,"name":"Capping Memory with cgroups v2 on an Oversubscribed VPS"}]},{"@type":"WebSite","@id":"https:\/\/virtualserversvps.com\/blog\/#website","url":"https:\/\/virtualserversvps.com\/blog\/","name":"Virtual Servers VPS Blog","description":"","potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/virtualserversvps.com\/blog\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"en-US"},{"@type":"Person","@id":"https:\/\/virtualserversvps.com\/blog\/#\/schema\/person\/82a299a8284a66ff49f97c74684724a0","name":"Virtual-Servers-Vps-Editor","image":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/virtualserversvps.com\/blog\/#\/schema\/person\/image\/","url":"https:\/\/secure.gravatar.com\/avatar\/d820b15f1cd028e97610d9adf536df7be5cb6423869967037d468d5355fa003f?s=96&d=mm&r=g","contentUrl":"https:\/\/secure.gravatar.com\/avatar\/d820b15f1cd028e97610d9adf536df7be5cb6423869967037d468d5355fa003f?s=96&d=mm&r=g","caption":"Virtual-Servers-Vps-Editor"},"sameAs":["https:\/\/virtualserversvps.com\/blog"],"url":"https:\/\/virtualserversvps.com\/blog\/author\/virtualserversvps\/"}]}},"_links":{"self":[{"href":"https:\/\/virtualserversvps.com\/blog\/wp-json\/wp\/v2\/posts\/1188","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/virtualserversvps.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/virtualserversvps.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/virtualserversvps.com\/blog\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/virtualserversvps.com\/blog\/wp-json\/wp\/v2\/comments?post=1188"}],"version-history":[{"count":1,"href":"https:\/\/virtualserversvps.com\/blog\/wp-json\/wp\/v2\/posts\/1188\/revisions"}],"predecessor-version":[{"id":1191,"href":"https:\/\/virtualserversvps.com\/blog\/wp-json\/wp\/v2\/posts\/1188\/revisions\/1191"}],"wp:attachment":[{"href":"https:\/\/virtualserversvps.com\/blog\/wp-json\/wp\/v2\/media?parent=1188"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/virtualserversvps.com\/blog\/wp-json\/wp\/v2\/categories?post=1188"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/virtualserversvps.com\/blog\/wp-json\/wp\/v2\/tags?post=1188"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}