{"id":1691,"date":"2026-11-15T22:39:46","date_gmt":"2026-11-15T22:39:46","guid":{"rendered":"https:\/\/bestdedicatedwebhostingserver.com\/blog\/?p=1691"},"modified":"2026-11-15T22:39:46","modified_gmt":"2026-11-15T22:39:46","slug":"hpa-keda-requests-limits-single-vps-side-project","status":"publish","type":"post","link":"https:\/\/bestdedicatedwebhostingserver.com\/blog\/hpa-keda-requests-limits-single-vps-side-project\/","title":{"rendered":"Kubernetes on Metal or a Single VPS: Request\/Limit Math for Small Game Backends"},"content":{"rendered":"\n<p class=\"wp-block-paragraph\">Kubernetes solves a problem most small game backends do not have: running many services across many machines with independent scaling and rolling deploys. If you run one session service, one matchmaker, and one database for a few hundred to a few thousand concurrent players, the fixed overhead of a control plane \u2014 etcd, API server, scheduler, controller manager, kubelet, CNI, ingress \u2014 is often larger than the workload it manages. This article does the arithmetic: what HPA and KEDA actually cost, how requests and limits interact, and at what player concurrency orchestrating becomes the cheaper option.<\/p>\n\n\n<!-- wp:heading level=\"2\" -->\n<h2>What the Control Plane Costs Before You Run Anything<\/h2>\n<!-- \/wp:post-content -->\n\n<!-- wp:paragraph -->\n<p>A k3s or kubeadm single-node install idles at a surprisingly specific footprint:<\/p>\n<!-- \/wp:paragraph -->\n\n<!-- wp:table -->\n<figure class=\"wp-block-table\"><table><thead><tr><th>Component<\/th><th>Idle CPU (cores)<\/th><th>Idle RAM<\/th><th>Disk<\/th><\/tr><\/thead><tbody><tr><td>etcd (single node)<\/td><td>0.05\u20130.15<\/td><td>150\u2013300 MB<\/td><td>1\u20134 GB, fsync-sensitive<\/td><\/tr><tr><td>kube-apiserver<\/td><td>0.05\u20130.20<\/td><td>250\u2013500 MB<\/td><td>\u2014<\/td><\/tr><tr><td>controller-manager + scheduler<\/td><td>0.05\u20130.10<\/td><td>150\u2013300 MB<\/td><td>\u2014<\/td><\/tr><tr><td>kubelet + containerd<\/td><td>0.05\u20130.15<\/td><td>200\u2013400 MB<\/td><td>\u2014<\/td><\/tr><tr><td>CNI + kube-proxy\/IPVS<\/td><td>0.02\u20130.10<\/td><td>50\u2013150 MB<\/td><td>\u2014<\/td><\/tr><tr><td>ingress controller<\/td><td>0.05\u20130.30<\/td><td>150\u2013400 MB<\/td><td>\u2014<\/td><\/tr><tr><td>metrics-server (for HPA)<\/td><td>0.02\u20130.10<\/td><td>50\u2013120 MB<\/td><td>\u2014<\/td><\/tr><tr><td><strong>Total idle<\/strong><\/td><td><strong>0.3\u20131.1<\/strong><\/td><td><strong>1.0\u20132.2 GB<\/strong><\/td><td><strong>~5 GB<\/strong><\/td><\/tr><\/tbody><\/table><figcaption class=\"wp-element-caption\">Single-node k3s with Flannel and ingress; a three-node control plane triples the etcd and API server figures and adds inter-node traffic.<\/figcaption><\/figure>\n<!-- \/wp:table -->\n\n<!-- wp:paragraph -->\n<p>KEDA adds an operator, a metrics adapter, and a scaler pod per <code>ScaledObject<\/code> class: budget another 0.15\u20130.3 cores and 300\u2013600 MB. On a 4-core\/16 GB box the total is roughly 25% of CPU and 15% of RAM before a single player connects; on 8 cores it is 10\u201314% of CPU. That tax is paid continuously, whether or not anyone is playing.<\/p>\n<!-- \/wp:paragraph -->\n\n<!-- wp:heading level=\"2\" -->\n<h2>Requests, Limits, and the Ratios That Matter<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:list -->\n<ul><li><strong>Requests<\/strong> are what the scheduler reserves. They do not cap usage. They determine bin-packing and the HPA&#8217;s utilisation denominator.<\/li><li><strong>Limits<\/strong> are enforced at runtime: CPU gets throttled via cfs_quota, memory gets the container OOM-killed.<\/li><\/ul>\n<!-- \/wp:list -->\n\n<!-- wp:paragraph -->\n<p>The asymmetry is the crux: a CPU limit throttles, a memory limit kills. Under-provision memory by 10% and you get a restart loop; under-provision CPU by 10% and you get tail latency. Be generous with memory limits relative to measured peak, and relatively tight with CPU requests.<\/p>\n<!-- \/wp:paragraph -->\n\n<!-- wp:table -->\n<figure class=\"wp-block-table\"><table><thead><tr><th>Ratio (limit \u00f7 request)<\/th><th>Effect<\/th><th>Use when<\/th><\/tr><\/thead><tbody><tr><td>1.0 (Guaranteed)<\/td><td>No throttling headroom; predictable<\/td><td>Latency-critical game session pods<\/td><\/tr><tr><td>1.2\u20131.5<\/td><td>Small burst absorbed, mild throttling possible<\/td><td>Most steady-state services<\/td><\/tr><tr><td>2.0\u20133.0<\/td><td>Significant burst capacity, throttling under load<\/td><td>Batch workers, cold-start-tolerant jobs<\/td><\/tr><tr><td>&gt; 3.0 (Burstable-heavy)<\/td><td>Node oversubscribed; noisy neighbours<\/td><td>Rarely justified for game backends<\/td><\/tr><\/tbody><\/table><\/figure>\n<!-- \/wp:table -->\n\n<!-- wp:paragraph -->\n<p>For a game session service, start with CPU request 0.5 and limit 0.75 per pod, memory request 512 Mi and limit 1 Gi. Measure the p99 CPU throttling ratio (<code>container_cpu_cfs_throttled_periods_total \/ container_cpu_cfs_periods_total<\/code>): anything above 5% under normal load means the limit is too low or the request is understated and the HPA denominator is wrong.<\/p>\n<!-- \/wp:paragraph -->\n\n<!-- wp:heading level=\"2\" -->\n<h2>The HPA Denominator Problem<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>HPA computes utilisation as <em>current usage \u00f7 requested<\/em>. That means the replica count it produces depends entirely on the request value, not on what the pod actually needs. Set a request of 0.2 cores for a pod that reliably consumes 0.6 and HPA sees 300% utilisation and over-scales. Set it at 1.0 for the same pod and HPA sees 60% and never scales, even while the pod is thrashing.<\/p>\n<!-- \/wp:paragraph -->\n<!-- wp:paragraph -->\n<p>HPA&#8217;s default <code>scaleUp<\/code> behaviour is also slower than a game backend wants: it checks every 15 s and then waits for new pods to become ready \u2014 2\u20135 seconds for a Go session service, 20\u201340 seconds for a JVM one including the startup probe. A 30-second spike is largely over by the time the extra replicas are serving.<\/p>\n<!-- \/wp:paragraph -->\n\n<!-- wp:heading level=\"2\" -->\n<h2>Where KEDA Genuinely Helps<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>HPA scales on resource utilisation. KEDA scales on external metrics \u2014 a queue length, a Kafka consumer lag, a Redis list depth, or a Prometheus query. For game backends, the honest cases where KEDA earns its overhead:<\/p>\n<!-- \/wp:paragraph -->\n\n<!-- wp:list -->\n<ul><li><strong>Queue-driven matchmaking workers<\/strong> \u2014 scale on queue depth rather than CPU, because the workers have known per-job cost.<\/li><li><strong>Post-session analytics<\/strong> \u2014 bursts that must drain quickly, scale to zero between bursts. Scale-to-zero is KEDA&#8217;s real advantage; HPA cannot go below one replica.<\/li><li><strong>Scheduled warm-up<\/strong> \u2014 KEDA&#8217;s cron scaler can pre-scale for a known event time, so pods are ready before the spike rather than 30 seconds into it.<\/li><\/ul>\n<!-- \/wp:list -->\n\n<!-- wp:paragraph -->\n<p>What KEDA does not fix: a bottleneck that is a shared database, a single-threaded game loop, or a connection pool. Scaling pods past the pool&#8217;s capacity produces contention, not throughput.<\/p>\n<!-- \/wp:paragraph -->\n\n<!-- wp:heading level=\"2\" -->\n<h2>Node Allocatable Math<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>The 4 cores on your box are not 4 cores of capacity. Reserve system, kubelet, and eviction headroom:<\/p>\n<!-- \/wp:paragraph -->\n<!-- wp:list -->\n<ul><li><strong>kube-reserved + system-reserved:<\/strong> typically 0.15 cores and 400 Mi on a 4-core node, scaling up with size.<\/li><li><strong>evictionHard memory:<\/strong> 100\u2013200 Mi held back.<\/li><li><strong>DaemonSets<\/strong> (CNI, kube-proxy, node exporter): 0.1\u20130.2 cores and 200\u2013400 Mi.<\/li><li><strong>Unallocated slack<\/strong> you should keep: 10\u201315% of remaining capacity, to avoid a node at 100% request where nothing can be scheduled or evicted cleanly.<\/li><\/ul>\n<!-- \/wp:list -->\n\n<!-- wp:table -->\n<figure class=\"wp-block-table\"><table><thead><tr><th>Node<\/th><th>Raw<\/th><th>Allocatable<\/th><th>After 12% slack<\/th><th>Usable for app pods<\/th><\/tr><\/thead><tbody><tr><td>4 vCPU \/ 8 GB<\/td><td>4 \/ 8192 Mi<\/td><td>3.85 \/ 7.4 Gi<\/td><td>3.39 \/ 6.5 Gi<\/td><td>~3.0 \/ 5.5 Gi<\/td><\/tr><tr><td>8 vCPU \/ 32 GB<\/td><td>8 \/ 32768 Mi<\/td><td>7.7 \/ 31.4 Gi<\/td><td>6.78 \/ 27.6 Gi<\/td><td>~6.3 \/ 26 Gi<\/td><\/tr><tr><td>16 vCPU \/ 64 GB<\/td><td>16 \/ 65536 Mi<\/td><td>15.6 \/ 63.2 Gi<\/td><td>13.7 \/ 55.6 Gi<\/td><td>~13 \/ 53 Gi<\/td><\/tr><\/tbody><\/table><\/figure>\n<!-- \/wp:table -->\n\n<!-- wp:paragraph -->\n<p>On the 4 vCPU node, after the control plane&#8217;s 0.3\u20131.1 cores, you have roughly 2 cores for application pods. That is two to four session-service replicas at the 0.5\u20130.75 request above. The box that looked like a 4-core machine is functionally a 2-core machine with a management layer.<\/p>\n<!-- \/wp:paragraph -->\n\n<!-- wp:heading level=\"2\" -->\n<h2>Memory Limits and the OOMKill Risk<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>Container memory accounting includes page cache that the container touched, which is why a pod can be killed long before its heap looks full. Two mitigations that matter:<\/p>\n<!-- \/wp:paragraph -->\n<!-- wp:list -->\n<ul><li><strong>Set the heap ceiling below the limit.<\/strong> For a JVM, <code>-XX:MaxRAMPercentage=70<\/code> with a 1 Gi limit leaves 300 MB for metaspace, thread stacks, and off-heap buffers. For Go, set <code>GOGC<\/code> and <code>GOMEMLIMIT<\/code> to ~80% of the limit so the runtime collects before the kernel intervenes.<\/li><li><strong>Never set a memory limit equal to the request unless you also set the request correctly.<\/strong> Guaranteed QoS is good for latency but leaves no flexibility; Burstable with request 70% and limit 100% is usually the right shape for a session service.<\/li><\/ul>\n<!-- \/wp:list -->\n\n<!-- wp:paragraph -->\n<p>Load-test to the limit and watch <code>container_memory_working_set_bytes<\/code>, not <code>container_memory_usage_bytes<\/code> \u2014 the former excludes reclaimable cache and is what the eviction decision uses. If working set at p99 peak is more than 85% of the limit, raise the limit.<\/p>\n<!-- \/wp:paragraph -->\n\n<!-- wp:heading level=\"2\" -->\n<h2>Cost Comparison: Three Deployments of the Same Backend<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:table -->\n<figure class=\"wp-block-table\"><table><thead><tr><th>Deployment<\/th><th>Monthly cost<\/th><th>Idle overhead<\/th><th>Time to add capacity<\/th><th>Failure isolation<\/th><\/tr><\/thead><tbody><tr><td>Single dedicated box, systemd units<\/td><td>$90\u2013$150<\/td><td>~0<\/td><td>Hours (provisioning)<\/td><td>None \u2014 one box, one blast radius<\/td><\/tr><tr><td>Single box, k3s + HPA + KEDA<\/td><td>$90\u2013$150 + 0.3 cores, 1.5 GB<\/td><td>10\u201325% CPU<\/td><td>Seconds (pods), hours (nodes)<\/td><td>Per-pod; box failure still total<\/td><\/tr><tr><td>3-node managed K8s cluster<\/td><td>$450\u2013$1,200<\/td><td>2 nodes&#8217; worth of control plane<\/td><td>Minutes (autoscaler)<\/td><td>Real \u2014 node failure tolerated<\/td><\/tr><\/tbody><\/table><\/figure>\n<!-- \/wp:table -->\n\n<!-- wp:paragraph -->\n<p>The single-box K8s option costs the same in dollars and 10\u201325% of its capacity in overhead, buying faster deploys, per-pod isolation, and a scale-out path that does not require re-architecting. Whether that is worth a quarter of a box is a judgment call \u2014 but it is not free.<\/p>\n<!-- \/wp:paragraph -->\n\n<!-- wp:heading level=\"2\" -->\n<h2>Decision Thresholds by Concurrency<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:paragraph -->\n<p>These are starting points, not laws \u2014 the shape of the workload matters more than the count, so assume a session service with roughly 10\u201320 KB of state per player and a state write every 15\u201330 seconds.<\/p>\n<!-- \/wp:paragraph -->\n<!-- wp:list -->\n<ul><li><strong>Under 200 concurrent players:<\/strong> one dedicated box, systemd or Compose, no orchestrator. A deploy is <code>systemctl restart<\/code> and a 5-second blip.<\/li><li><strong>200\u20131,000:<\/strong> one bigger box, or two with a load balancer and session affinity, if you need per-service isolation or zero-downtime deploys. If you already know Kubernetes well, k3s here is defensible; if you are learning it, learn in staging.<\/li><li><strong>1,000\u20135,000:<\/strong> multi-node orchestration starts earning its cost. Two or three game-session nodes plus managed database, HPA on CPU or connection count, KEDA for any queue-driven component.<\/li><li><strong>Above 5,000 across regions:<\/strong> orchestration is the only practical option; the hard problem becomes placement and cross-region state, not pod sizing.<\/li><\/ul>\n<!-- \/wp:list -->\n\n<!-- wp:paragraph -->\n<p>Between 200 and 1,000 players the deciding factor is team familiarity rather than mathematics: the break-even is really the point where distinct services and deploy targets exceed what one shell script can manage.<\/p>\n<!-- \/wp:paragraph -->\n\n<!-- wp:heading level=\"2\" -->\n<h2>If You Do Run It on One Box: Non-Negotiable Settings<\/h2>\n<!-- \/wp:heading -->\n\n<!-- wp:list -->\n<ul><li><strong>Reserve the control plane&#8217;s memory<\/strong> with <code>system-reserved<\/code>, or the kubelet and your game pods will compete during a GC pause.<\/li><li><strong>etcd on NVMe, fsync enabled.<\/strong> etcd fsync latency above 10 ms produces API server timeouts and flapping controllers \u2014 a failure mode that looks like a networking bug.<\/li><li><strong><code>podDisruptionBudget<\/code> on the session service<\/strong> even with one replica, so an upgrade does not take everything down at once.<\/li><li><strong>Startup and readiness probes that reflect real readiness<\/strong>, not just &#8220;port open&#8221;. A session pod that accepts connections before it has loaded state will drop the first players that reach it.<\/li><li><strong>Limit range on the namespace<\/strong> so a pod without a limit cannot be deployed \u2014 an unlimited pod on a single box can consume all memory and take the control plane with it.<\/li><\/ul>\n<!-- \/wp:list -->\n\n<!-- wp:paragraph -->\n<p>None of this argues against Kubernetes \u2014 it argues for pricing it accurately. Set requests from measured p99 usage rather than a round number guessed at deploy time; guess round numbers and HPA will scale on a fiction.<\/p>\n<!-- \/wp:paragraph -->\n","protected":false},"excerpt":{"rendered":"<p>Kubernetes solves a problem most small game backends do not have: running many services across many machines with independent scaling and rolling deploys. If you run one session service, one&#8230;<\/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":[17],"tags":[],"class_list":["post-1691","post","type-post","status-publish","format-standard","hentry","category-dedicated-hosting-tutorials"],"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>Kubernetes on Metal or a Single VPS: Request\/Limit Math for Small Game Backends - Best Dedicated Web Hosting Server Blog<\/title>\n<meta name=\"description\" content=\"Kubernetes solves a problem most small game backends do not have: running many services across many machines with independent scaling and rolling deploys.\" \/>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/bestdedicatedwebhostingserver.com\/blog\/hpa-keda-requests-limits-single-vps-side-project\/\" \/>\n<meta property=\"og:locale\" content=\"en_US\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"Kubernetes on Metal or a Single VPS: Request\/Limit Math for Small Game Backends\" \/>\n<meta property=\"og:description\" content=\"Kubernetes on Metal or a Single VPS: Request\/Limit Math for Small Game Backends\" \/>\n<meta property=\"og:url\" content=\"https:\/\/bestdedicatedwebhostingserver.com\/blog\/hpa-keda-requests-limits-single-vps-side-project\/\" \/>\n<meta property=\"og:site_name\" content=\"Best Dedicated Web Hosting Server Blog\" \/>\n<meta property=\"article:published_time\" content=\"2026-11-15T22:39:46+00:00\" \/>\n<meta name=\"author\" content=\"bestdedicatedwebhostingserver\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:creator\" content=\"@bestdedicatedwebhostingserver\" \/>\n<meta name=\"twitter:label1\" content=\"Written by\" \/>\n\t<meta name=\"twitter:data1\" content=\"bestdedicatedwebhostingserver\" \/>\n\t<meta name=\"twitter:label2\" content=\"Est. reading time\" \/>\n\t<meta name=\"twitter:data2\" content=\"7 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"WebPage\",\"@id\":\"https:\/\/bestdedicatedwebhostingserver.com\/blog\/hpa-keda-requests-limits-single-vps-side-project\/\",\"url\":\"https:\/\/bestdedicatedwebhostingserver.com\/blog\/hpa-keda-requests-limits-single-vps-side-project\/\",\"name\":\"Kubernetes on Metal or a Single VPS: Request\/Limit Math for Small Game Backends - Best Dedicated Web Hosting Server Blog\",\"isPartOf\":{\"@id\":\"https:\/\/bestdedicatedwebhostingserver.com\/blog\/#website\"},\"datePublished\":\"2026-11-15T22:39:46+00:00\",\"author\":{\"@id\":\"https:\/\/bestdedicatedwebhostingserver.com\/blog\/#\/schema\/person\/6eb5f9ea35033fe8e67df44397e089b1\"},\"description\":\"Kubernetes solves a problem most small game backends do not have: running many services across many machines with independent scaling and rolling deploys.\",\"breadcrumb\":{\"@id\":\"https:\/\/bestdedicatedwebhostingserver.com\/blog\/hpa-keda-requests-limits-single-vps-side-project\/#breadcrumb\"},\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/bestdedicatedwebhostingserver.com\/blog\/hpa-keda-requests-limits-single-vps-side-project\/\"]}]},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/bestdedicatedwebhostingserver.com\/blog\/hpa-keda-requests-limits-single-vps-side-project\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\/\/bestdedicatedwebhostingserver.com\/blog\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"Kubernetes on Metal or a Single VPS: Request\/Limit Math for Small Game Backends\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/bestdedicatedwebhostingserver.com\/blog\/#website\",\"url\":\"https:\/\/bestdedicatedwebhostingserver.com\/blog\/\",\"name\":\"Best Dedicated Web Hosting Server Blog\",\"description\":\"\",\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/bestdedicatedwebhostingserver.com\/blog\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"en-US\"},{\"@type\":\"Person\",\"@id\":\"https:\/\/bestdedicatedwebhostingserver.com\/blog\/#\/schema\/person\/6eb5f9ea35033fe8e67df44397e089b1\",\"name\":\"bestdedicatedwebhostingserver\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\/\/bestdedicatedwebhostingserver.com\/blog\/#\/schema\/person\/image\/\",\"url\":\"https:\/\/secure.gravatar.com\/avatar\/462a07aa37399bdb149e5f91dc3cd8906656bc4c7ed391a3b6f0199c5d2ab964?s=96&d=mm&r=g\",\"contentUrl\":\"https:\/\/secure.gravatar.com\/avatar\/462a07aa37399bdb149e5f91dc3cd8906656bc4c7ed391a3b6f0199c5d2ab964?s=96&d=mm&r=g\",\"caption\":\"bestdedicatedwebhostingserver\"},\"sameAs\":[\"https:\/\/bestdedicatedwebhostingserver.com\/blog\",\"https:\/\/x.com\/bestdedicatedwebhostingserver\"],\"url\":\"https:\/\/bestdedicatedwebhostingserver.com\/blog\/author\/bestdedicatedwebhostingserver\/\"}]}<\/script>\n<!-- \/ Yoast SEO Premium plugin. -->","yoast_head_json":{"title":"Kubernetes on Metal or a Single VPS: Request\/Limit Math for Small Game Backends - Best Dedicated Web Hosting Server Blog","description":"Kubernetes solves a problem most small game backends do not have: running many services across many machines with independent scaling and rolling deploys.","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:\/\/bestdedicatedwebhostingserver.com\/blog\/hpa-keda-requests-limits-single-vps-side-project\/","og_locale":"en_US","og_type":"article","og_title":"Kubernetes on Metal or a Single VPS: Request\/Limit Math for Small Game Backends","og_description":"Kubernetes on Metal or a Single VPS: Request\/Limit Math for Small Game Backends","og_url":"https:\/\/bestdedicatedwebhostingserver.com\/blog\/hpa-keda-requests-limits-single-vps-side-project\/","og_site_name":"Best Dedicated Web Hosting Server Blog","article_published_time":"2026-11-15T22:39:46+00:00","author":"bestdedicatedwebhostingserver","twitter_card":"summary_large_image","twitter_creator":"@bestdedicatedwebhostingserver","twitter_misc":{"Written by":"bestdedicatedwebhostingserver","Est. reading time":"7 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"WebPage","@id":"https:\/\/bestdedicatedwebhostingserver.com\/blog\/hpa-keda-requests-limits-single-vps-side-project\/","url":"https:\/\/bestdedicatedwebhostingserver.com\/blog\/hpa-keda-requests-limits-single-vps-side-project\/","name":"Kubernetes on Metal or a Single VPS: Request\/Limit Math for Small Game Backends - Best Dedicated Web Hosting Server Blog","isPartOf":{"@id":"https:\/\/bestdedicatedwebhostingserver.com\/blog\/#website"},"datePublished":"2026-11-15T22:39:46+00:00","author":{"@id":"https:\/\/bestdedicatedwebhostingserver.com\/blog\/#\/schema\/person\/6eb5f9ea35033fe8e67df44397e089b1"},"description":"Kubernetes solves a problem most small game backends do not have: running many services across many machines with independent scaling and rolling deploys.","breadcrumb":{"@id":"https:\/\/bestdedicatedwebhostingserver.com\/blog\/hpa-keda-requests-limits-single-vps-side-project\/#breadcrumb"},"inLanguage":"en-US","potentialAction":[{"@type":"ReadAction","target":["https:\/\/bestdedicatedwebhostingserver.com\/blog\/hpa-keda-requests-limits-single-vps-side-project\/"]}]},{"@type":"BreadcrumbList","@id":"https:\/\/bestdedicatedwebhostingserver.com\/blog\/hpa-keda-requests-limits-single-vps-side-project\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/bestdedicatedwebhostingserver.com\/blog\/"},{"@type":"ListItem","position":2,"name":"Kubernetes on Metal or a Single VPS: Request\/Limit Math for Small Game Backends"}]},{"@type":"WebSite","@id":"https:\/\/bestdedicatedwebhostingserver.com\/blog\/#website","url":"https:\/\/bestdedicatedwebhostingserver.com\/blog\/","name":"Best Dedicated Web Hosting Server Blog","description":"","potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/bestdedicatedwebhostingserver.com\/blog\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"en-US"},{"@type":"Person","@id":"https:\/\/bestdedicatedwebhostingserver.com\/blog\/#\/schema\/person\/6eb5f9ea35033fe8e67df44397e089b1","name":"bestdedicatedwebhostingserver","image":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/bestdedicatedwebhostingserver.com\/blog\/#\/schema\/person\/image\/","url":"https:\/\/secure.gravatar.com\/avatar\/462a07aa37399bdb149e5f91dc3cd8906656bc4c7ed391a3b6f0199c5d2ab964?s=96&d=mm&r=g","contentUrl":"https:\/\/secure.gravatar.com\/avatar\/462a07aa37399bdb149e5f91dc3cd8906656bc4c7ed391a3b6f0199c5d2ab964?s=96&d=mm&r=g","caption":"bestdedicatedwebhostingserver"},"sameAs":["https:\/\/bestdedicatedwebhostingserver.com\/blog","https:\/\/x.com\/bestdedicatedwebhostingserver"],"url":"https:\/\/bestdedicatedwebhostingserver.com\/blog\/author\/bestdedicatedwebhostingserver\/"}]}},"_links":{"self":[{"href":"https:\/\/bestdedicatedwebhostingserver.com\/blog\/wp-json\/wp\/v2\/posts\/1691","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/bestdedicatedwebhostingserver.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/bestdedicatedwebhostingserver.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/bestdedicatedwebhostingserver.com\/blog\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/bestdedicatedwebhostingserver.com\/blog\/wp-json\/wp\/v2\/comments?post=1691"}],"version-history":[{"count":1,"href":"https:\/\/bestdedicatedwebhostingserver.com\/blog\/wp-json\/wp\/v2\/posts\/1691\/revisions"}],"predecessor-version":[{"id":1694,"href":"https:\/\/bestdedicatedwebhostingserver.com\/blog\/wp-json\/wp\/v2\/posts\/1691\/revisions\/1694"}],"wp:attachment":[{"href":"https:\/\/bestdedicatedwebhostingserver.com\/blog\/wp-json\/wp\/v2\/media?parent=1691"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/bestdedicatedwebhostingserver.com\/blog\/wp-json\/wp\/v2\/categories?post=1691"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/bestdedicatedwebhostingserver.com\/blog\/wp-json\/wp\/v2\/tags?post=1691"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}