Note
This is a fork of postfinance/kuota-calc that adds some features as the original doesn't seem to be maintained. But neither will this. After I added what I need, I will stop development as well. Feel free to fork as well :)
Simple utility to calculate the maximum needed resource quota for deployment(s). kuota-calc takes the deployment strategy, replicas and all containers into account, see supported-resources for a list of kubernetes resources which are currently supported by kuota-calc.
In shared environments such as kubernetes it is always a good idea to isolate/constrain different workloads to prevent them from interfering each other. Kubernetes provides Resource Quotas to limit compute, storage and object resources of namespaces.
Calculating the needed compute resources can be a bit challenging (especially with large and complex deployments) because we must respect certain settings/defaults like the deployment strategy, number of replicas and so on. This is where kuota-calc can help you, it calculates the maximum needed resource quota in order to be able to start a deployment of all resources at the same time by respecting deployment strategies, replicas and so on.
Get a detailed report of all resources, their max required quota and a total.
$ cat examples/deployment.yaml | kuota-calc -detailed
Version Kind Name Replicas Strategy MaxReplicas CPURequest CPULimit MemoryRequest MemoryLimit
apps/v1 Deployment myapp 10 RollingUpdate 13 3250m 6500m 832Mi 3328Mi
apps/v1 StatefulSet myapp 3 RollingUpdate 3 750m 3 6Gi 12Gi
Table and Total assuming simultaneous rollout of all resources
Total
CPU Request: 4
CPU Limit: 9500m
Memory Request: 6976Mi
Memory Limit: 15616MiFor comparison, here the simultaneous rollout is limited to zero resources, so you get the required quotas to just run, but not deploy the applications.
$ cat examples/deployment.yaml | kuota-calc --max-rollouts=0
CPU Request: 3250m
CPU Limit: 8
Memory Request: 6784Mi
Memory Limit: 14848MiTo calc usage for deploymentConfigs, deployments and statefulSets deployed in an openshift cluster:
$ oc get dc,sts,deploy -o json | yq -p=json -o=yaml '.items[] | split_doc' | kuota-calc --detailed
Warning: apps.openshift.io/v1 DeploymentConfig is deprecated in v4.14+, unavailable in v4.10000+
Version Kind Name Replicas Strategy MaxReplicas CPURequest CPULimit MemoryRequest MemoryLimit
apps.openshift.io/v1 DeploymentConfig my-app-1 0 Recreate 0 0 0 0 0
apps.openshift.io/v1 DeploymentConfig my-app-2 1 Recreate 1 1250m 1700m 500Mi 500Mi
apps/v1 StatefulSet my-app-3 1 RollingUpdate 1 150m 1100m 2200Mi 2200Mi
apps/v1 Deployment my-app-4 1 RollingUpdate 2 100m 200m 100Mi 512Mi
Total
CPU Request: 1500m
CPU Limit: 3
Memory Request: 2800Mi
Memory Limit: 3212MiPre-compiled statically linked binaries are available on the releases page.
kuota-calc can either be used as a kubectl plugin or invoked directly. If you intend to use kuota-calc as
a kubectl plugin, simply place the binary anywhere in $PATH named kubectl-kuota_calc with execute permissions.
For further information, see the official documentation on kubectl plugins here.
currently the kubectl plugin is not released for this fork
kuota-calc is still a work-in progress, there are plans to support more k8s resources (see #5 for more info).
Currently supported:
- apps.openshift.io/v1 DeploymentConfig
- apps/v1 Deployment
- apps/v1 StatefulSet
- apps/v1 DaemonSet
- batch/v1 CronJob
- batch/v1 Job
- v1 Pod
- CronJobs: the cron concurrencyPolicy is not considered, a CronJob is treated as a single Pod (#18)
- DaemonSet: neither node count nor UpdateStrategy are considered. Treated as a single Pod. (#21)