{"id":18684577,"url":"https://github.com/kubevirt/common-instancetypes","last_synced_at":"2025-04-12T04:32:16.934Z","repository":{"id":62552089,"uuid":"558738411","full_name":"kubevirt/common-instancetypes","owner":"kubevirt","description":"Instancetypes and preferences for running VMs on KubeVirt","archived":false,"fork":false,"pushed_at":"2025-04-10T07:02:03.000Z","size":15588,"stargazers_count":29,"open_issues_count":10,"forks_count":17,"subscribers_count":9,"default_branch":"main","last_synced_at":"2025-04-10T08:21:10.339Z","etag":null,"topics":["instance-type","kubernetes","kubevirt","preferences"],"latest_commit_sha":null,"homepage":"https://kubevirt.io/user-guide/virtual_machines/instancetypes/#common-instancetypes","language":"Go","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"apache-2.0","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/kubevirt.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2022-10-28T07:24:12.000Z","updated_at":"2025-04-08T12:50:07.000Z","dependencies_parsed_at":"2024-06-21T08:29:16.512Z","dependency_job_id":"7dc6dd65-f700-427d-82aa-4aad7298247b","html_url":"https://github.com/kubevirt/common-instancetypes","commit_stats":null,"previous_names":[],"tags_count":23,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/kubevirt%2Fcommon-instancetypes","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/kubevirt%2Fcommon-instancetypes/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/kubevirt%2Fcommon-instancetypes/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/kubevirt%2Fcommon-instancetypes/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/kubevirt","download_url":"https://codeload.github.com/kubevirt/common-instancetypes/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":248517284,"owners_count":21117425,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["instance-type","kubernetes","kubevirt","preferences"],"created_at":"2024-11-07T10:18:12.189Z","updated_at":"2025-04-12T04:32:11.924Z","avatar_url":"https://github.com/kubevirt.png","language":"Go","funding_links":[],"categories":[],"sub_categories":[],"readme":"# kubevirt/common-instancetypes\n\nA set of [instance types and preferences](https://kubevirt.io/user-guide/virtual_machines/instancetypes/) to help create [`KubeVirt`](https://kubevirt.io)  [`VirtualMachines`](http://kubevirt.io/api-reference/main/definitions.html#_v1alpha1_virtualmachine).\n\n# Overview\n\n## Structure\n\nThe available instance types can be divided into two categories:\n\n1. Workload agnostic - or general purpose\n2. Workload specific\n\nWorkload agnostic instance types are a good starting point to run your workload.\nOnce you know more about the requirements of your workload, you can start choosing a\nworkload specific instance type.\n\nThe following diagram summarizes the available instance types and their use-cases:\n\n```mermaid\ngraph TD\n\nclassDef grp fill:white,stroke:lightgray,color:gray\nclassDef series fill:lightyellow,stroke:lightgray\nclassDef instancetype fill:\n\nwrklda(Workload agnostic)\nwrklds(Workload specific)\nclass wrklda grp\n\nwrklda:::grp --\u003e Universal:::series\nUniversal([Universal]):::series --\u003e u1:::instancetype\n\nwrklda:::grp --\u003e Overcommitted:::series\nOvercommitted([Overcommitted]):::series --\u003e o1:::instancetype\n\nwrklds:::grp --\u003e Computeexclusive:::series\nComputeexclusive([Compute Exclusive]):::series --\u003e cx1:::instancetype\n\nwrklds:::grp --\u003e GPUNVIDIA:::series\nGPUNVIDIA([GPU NVIDIA]):::series --\u003e gn1:::instancetype\n\nwrklds:::grp --\u003e Memory:::series\nMemory([Memory]):::series --\u003e m1:::instancetype\n\nwrklds:::grp --\u003e Network:::series\nNetwork([Network]):::series --\u003e n1:::instancetype\n\nwrklds:::grp --\u003e Realtime:::series\nRealtime([Realtime]):::series --\u003e rt1:::instancetype\n```\n\n### Labels\n\nVarious labels are used to decorate objects provided by this project to help end users find the right resources for them. See [docs/labels.md](./docs/labels.md) for more details.\n\n### Schema\n\n\u003cdetails\u003e\u003csummary\u003eClick in order to view the instance type names schema\u003c/summary\u003e\n\n```ebnf\ninstanceTypeName = seriesName , \".\" , size;\n\nseriesName = ( class | vendorClass ) , version;\n\nclass = \"u\" | \"o\" | \"cx\" | \"g\" | \"m\" | \"n\" | \"rt\";\nvendorClass = \"g\" , vendorHint;\nvendorHint = \"n\" | \"i\" | \"a\";\nversion = \"1\";\n\nsize = \"small\" | \"medium\" | \"large\" | [( \"2\" | \"4\" | \"8\" )] , \"xlarge\";\n```\n\u003c/details\u003e\n\n# Series\n\n.                           |  U  |  O  |  CX  |  GN  |  M  |  N  |  RT\n----------------------------|-----|-----|------|------|-----|-----|------\n*Has GPUs*                  |     |     |      |  ✓   |     |     |\n*Hugepages*                 |     |     |  ✓   |      |  ✓  |  ✓  |  ✓\n*Overcommitted Memory*      |     |  ✓  |      |      |     |     |\n*Dedicated CPU*             |     |     |  ✓   |      |     |  ✓  |  ✓\n*Burstable CPU performance* |  ✓  |  ✓  |      |  ✓   |  ✓  |     |\n*Isolated emulator threads* |     |     |  ✓   |      |     |  ✓  |  ✓\n*vNUMA*                     |     |     |  ✓   |      |     |     |  ✓\n*vCPU-To-Memory Ratio*      | 1:4 | 1:4 |  1:2 |  1:4 | 1:8 | 1:2 | 1:4\n\n## U Series\n\nThe U Series is quite neutral and provides resources for\ngeneral purpose applications.\n\n*U* is the abbreviation for \"Universal\", hinting at the universal\nattitude towards workloads.\n\nVMs of instance types will share physical CPU cores on a\ntime-slice basis with other VMs.\n\n### U Series Characteristics\n\nSpecific characteristics of this series are:\n- *Burstable CPU performance* - The workload has a baseline compute\n  performance but is permitted to burst beyond this baseline, if\n  excess compute resources are available.\n- *vCPU-To-Memory Ratio (1:4)* - A vCPU-to-Memory ratio of 1:4, for less\n  noise per node.\n\n## O Series\n\nThe O Series is based on the U Series, with the only difference\nbeing that memory is overcommitted.\n\n*O* is the abbreviation for \"Overcommitted\".\n\n### UO Series Characteristics\n\nSpecific characteristics of this series are:\n- *Burstable CPU performance* - The workload has a baseline compute\n  performance but is permitted to burst beyond this baseline, if\n  excess compute resources are available.\n- *Overcommitted Memory* - Memory is over-committed in order to achieve\n  a higher workload density.\n- *vCPU-To-Memory Ratio (1:4)* - A vCPU-to-Memory ratio of 1:4, for less\n  noise per node.\n\n## CX Series\n\nThe CX Series provides exclusive compute resources for compute\nintensive applications.\n\n*CX* is the abbreviation of \"Compute Exclusive\".\n\nThe exclusive resources are given to the compute threads of the\nVM. In order to ensure this, some additional cores (depending\non the number of disks and NICs) will be requested to offload\nthe IO threading from cores dedicated to the workload.\nIn addition, in this series, the NUMA topology of the used\ncores is provided to the VM.\n\n### CX Series Characteristics\n\nSpecific characteristics of this series are:\n- *Hugepages* - Hugepages are used in order to improve memory\n  performance.\n- *Dedicated CPU* - Physical cores are exclusively assigned to every\n  vCPU in order to provide fixed and high compute guarantees to the\n  workload.\n- *Isolated emulator threads* - Hypervisor emulator threads are isolated\n  from the vCPUs in order to reduce emaulation related impact on the\n  workload.\n- *vNUMA* - Physical NUMA topology is reflected in the guest in order to\n  optimize guest sided cache utilization.\n- *vCPU-To-Memory Ratio (1:2)* - A vCPU-to-Memory ratio of 1:2.\n\n## GN Series\n\nThe GN Series provides instances types intended for VMs with\nNVIDIA GPU resources attached.\n\n*GN* is the abbreviation of \"GPU NVIDIA\".\n\nThis series is intended to be used with VMs consuming GPUs\nprovided by the\n[NVIDIA GPU Operator](https://github.com/NVIDIA/gpu-operator)\nwhich can be installed on Kubernetes and also is made available\non OpenShift via OperatorHub.\n\n### GN Series Characteristics\n\nSpecific characteristics of this series are:\n- *Has GPUs* - Has GPUs predefined.\n- *Burstable CPU performance* - The workload has a baseline compute\n  performance but is permitted to burst beyond this baseline, if\n  excess compute resources are available.\n- *vCPU-To-Memory Ratio (1:4)* - A vCPU-to-Memory ratio of 1:4, for less\n  noise per node.\n\n## M Series\n\nThe M Series provides resources for memory intensive\napplications.\n\n*M* is the abbreviation of \"Memory\".\n\n### M Series Characteristics\n\nSpecific characteristics of this series are:\n- *Hugepages* - Hugepages are used in order to improve memory\n  performance.\n- *Burstable CPU performance* - The workload has a baseline compute\n  performance but is permitted to burst beyond this baseline, if\n  excess compute resources are available.\n- *vCPU-To-Memory Ratio (1:8)* - A vCPU-to-Memory ratio of 1:8, for much\n  less noise per node.\n\n## N Series\n\nThe N Series provides resources for network intensive DPDK\napplications, like VNFs.\n\n*N* is the abbreviation for \"Network\".\n\nThis series of instancetypes requires nodes capable\nof running DPDK workloads and being marked with the respective\nnode-role.kubevirt.io/worker-dpdk label as such.\n\n### N Series Characteristics\n\nSpecific characteristics of this series are:\n- *Hugepages* - Hugepages are used in order to improve memory\n  performance.\n- *Dedicated CPU* - Physical cores are exclusively assigned to every\n  vCPU in order to provide fixed and high compute guarantees to the\n  workload.\n- *Isolated emulator threads* - Hypervisor emulator threads are isolated\n  from the vCPUs in order to reduce emaulation related impact on the\n  workload.\n- *vCPU-To-Memory Ratio (1:2)* - A vCPU-to-Memory ratio of 1:2.\n\n## RT Series\n\nThe RT Series provides resources for realtime applications, like Oslat.\n\n*RT* is the abbreviation for \"realtime\".\n\nThis series of instance types requires nodes capable of running\nrealtime applications.\n\n### RT Series Characteristics\n\nSpecific characteristics of this series are:\n- *Hugepages* - Hugepages are used in order to improve memory\n  performance.\n- *Dedicated CPU* - Physical cores are exclusively assigned to every\n  vCPU in order to provide fixed and high compute guarantees to the\n  workload.\n- *Isolated emulator threads* - Hypervisor emulator threads are isolated\n  from the vCPUs in order to reduce emaulation related impact on the\n  workload.\n- *vCPU-To-Memory Ratio (1:4)* - A vCPU-to-Memory ratio of 1:4 starting from\n  the medium size.\n\n## Development\n\nTo get started with customizing or creating your own instancetypes and preferences\nsee [DEVELOPMENT.md](./DEVELOPMENT.md).\n\n## Resources\n\nThe following instancetype resources (cluster-wide and namespaced) are\nprovided by this project:\n\nName | vCPUs | Memory\n-----|-------|-------\ncx1.2xlarge  |  8  |  16Gi\ncx1.4xlarge  |  16  |  32Gi\ncx1.8xlarge  |  32  |  64Gi\ncx1.large  |  2  |  4Gi\ncx1.medium  |  1  |  2Gi\ncx1.xlarge  |  4  |  8Gi\ngn1.2xlarge  |  8  |  32Gi\ngn1.4xlarge  |  16  |  64Gi\ngn1.8xlarge  |  32  |  128Gi\ngn1.xlarge  |  4  |  16Gi\nm1.2xlarge  |  8  |  64Gi\nm1.4xlarge  |  16  |  128Gi\nm1.8xlarge  |  32  |  256Gi\nm1.large  |  2  |  16Gi\nm1.xlarge  |  4  |  32Gi\nn1.2xlarge  |  16  |  32Gi\nn1.4xlarge  |  32  |  64Gi\nn1.8xlarge  |  64  |  128Gi\nn1.large  |  4  |  8Gi\nn1.medium  |  4  |  4Gi\nn1.xlarge  |  8  |  16Gi\no1.2xlarge  |  8  |  32Gi\no1.4xlarge  |  16  |  64Gi\no1.8xlarge  |  32  |  128Gi\no1.large  |  2  |  8Gi\no1.medium  |  1  |  4Gi\no1.micro  |  1  |  1Gi\no1.nano  |  1  |  512Mi\no1.small  |  1  |  2Gi\no1.xlarge  |  4  |  16Gi\nrt1.2xlarge  |  8  |  32Gi\nrt1.4xlarge  |  16  |  64Gi\nrt1.8xlarge  |  32  |  128Gi\nrt1.large  |  2  |  8Gi\nrt1.medium  |  1  |  4Gi\nrt1.micro  |  1  |  1Gi\nrt1.small  |  1  |  2Gi\nrt1.xlarge  |  4  |  16Gi\nu1.2xlarge  |  8  |  32Gi\nu1.2xmedium  |  2  |  4Gi\nu1.4xlarge  |  16  |  64Gi\nu1.8xlarge  |  32  |  128Gi\nu1.large  |  2  |  8Gi\nu1.medium  |  1  |  4Gi\nu1.micro  |  1  |  1Gi\nu1.nano  |  1  |  512Mi\nu1.small  |  1  |  2Gi\nu1.xlarge  |  4  |  16Gi\n\nThe following preference resources (cluster-wide and namespaced) are\nprovided by this project:\n\nName | Guest OS\n-----|---------\nalpine | Alpine\ncentos.7 | CentOS 7\ncentos.7.desktop | CentOS 7\ncentos.stream10 | CentOS Stream 10\ncentos.stream10.desktop | CentOS Stream 10\ncentos.stream8 | CentOS Stream 8\ncentos.stream8.desktop | CentOS Stream 8\ncentos.stream8.dpdk | CentOS Stream 8\ncentos.stream9 | CentOS Stream 9\ncentos.stream9.desktop | CentOS Stream 9\ncentos.stream9.dpdk | CentOS Stream 9\ncirros | Cirros\nfedora | Fedora (amd64)\nfedora.arm64 | Fedora (arm64)\nopensuse.leap | OpenSUSE Leap\nopensuse.tumbleweed | OpenSUSE Tumbleweed\nrhel.10 | Red Hat Enterprise Linux 10 Beta (amd64)\nrhel.10.arm64 | Red Hat Enterprise Linux 10 Beta (arm64)\nrhel.7 | Red Hat Enterprise Linux 7\nrhel.7.desktop | Red Hat Enterprise Linux 7\nrhel.8 | Red Hat Enterprise Linux 8\nrhel.8.desktop | Red Hat Enterprise Linux 8\nrhel.8.dpdk | Red Hat Enterprise Linux 8\nrhel.9 | Red Hat Enterprise Linux 9 (amd64)\nrhel.9.arm64 | Red Hat Enterprise Linux 9 (arm64)\nrhel.9.desktop | Red Hat Enterprise Linux 9 Desktop (amd64)\nrhel.9.dpdk | Red Hat Enterprise Linux 9 DPDK (amd64)\nrhel.9.realtime | Red Hat Enterprise Linux 9 Realtime (amd64)\nsles | SUSE Linux Enterprise Server\nubuntu | Ubuntu\nwindows.10 | Microsoft Windows 10\nwindows.10.virtio | Microsoft Windows 10 (virtio)\nwindows.11 | Microsoft Windows 11\nwindows.11.virtio | Microsoft Windows 11 (virtio)\nwindows.2k16 | Microsoft Windows Server 2016\nwindows.2k16.virtio | Microsoft Windows Server 2016 (virtio)\nwindows.2k19 | Microsoft Windows Server 2019\nwindows.2k19.virtio | Microsoft Windows Server 2019 (virtio)\nwindows.2k22 | Microsoft Windows Server 2022\nwindows.2k22.virtio | Microsoft Windows Server 2022 (virtio)\nwindows.2k25 | Microsoft Windows Server 2025\nwindows.2k25.virtio | Microsoft Windows Server 2025 (virtio)\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fkubevirt%2Fcommon-instancetypes","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fkubevirt%2Fcommon-instancetypes","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fkubevirt%2Fcommon-instancetypes/lists"}