OsUpdateDeploymentResources¶
Updates container resource requests and limits in a Kubernetes or OpenShift deployment.
What It Does¶
Updates CPU and memory requests and limits on a deployment container, then waits for the deployment to converge.
This is useful when a scenario needs a temporary resource profile, for example to test scaling thresholds or resource-constrained behavior.
YAML Example¶
Sessions:
- Name: ProbeSession
Probes:
- Name: OsUpdateDeploymentResourcesProbe
Probe: OsUpdateDeploymentResources
ProbeConfiguration:
UseGlobalDict: true
ReplicaSetName: orders-api
ContainerName: api
IntervalBetweenDesiredStateChecksMs: 1000
TimeoutWaitForDesiredStateSeconds: 300
Openshift:
Cluster: https://api.cluster.local:6443
Namespace: docs
Username: docs-user
Password: docs-password
DesiredResources:
Limits:
Cpu: 1000m
Memory: 1Gi
Requests:
Cpu: 250m
Memory: 256Mi
What This Configuration Does¶
This probe updates the api container in the orders-api deployment so that it requests 250m CPU and 256Mi memory, with limits of 1000m CPU and 1Gi memory.
The deployment is then allowed to roll out and settle before the scenario continues.
Global Dictionary Behavior¶
With UseGlobalDict: true, missing shared cluster settings can be resolved from Os/Defaults, and missing DesiredResources can be restored from Os/Recovery/Resources/Deployment/<ReplicaSetName>/<ContainerName> after an earlier probe in the same execution and session captured the pre-change state.
The probe writes its pre-change snapshot to the unique canonical scoped path for the current probe execution and then updates the recovery alias so a later rollback probe can reuse it. This is useful when you want to test a temporary CPU or memory profile and then restore the original requests and limits.
No additional per-probe recovery caveat applies beyond the execution and session scoping rules.
When UseGlobalDict is false, the probe keeps the current behavior: it uses only local YAML or code configuration and does not read or write probe-global-dictionary state.