Scaling and Availability Configuration

When CodeMie API is deployed on Kubernetes, the Helm charts provide two built-in mechanisms to control how many pods run and how disruptions to those pods are handled:

• Horizontal Pod Autoscaler (HPA) — automatically adjusts the number of running pods up or down based on observed CPU utilization. Use it to handle variable load without manually changing replica counts.
• Pod Disruption Budget (PDB) — limits how many pods can be taken down at the same time during voluntary disruptions, such as node draining, cluster upgrades, etc. Use it to keep a minimum level of service available while the underlying infrastructure changes.

Both are disabled by default and can be enabled and tuned through the CodeMie API Helm chart values.

Horizontal Pod Autoscaler (HPA)

When enabled, the HPA watches the CPU utilization of the CodeMie API pods and scales the number of replicas between a configured minimum and maximum to keep utilization near a target percentage.

Configuration

Set the hpa block in your values-<cloud>.yaml:

hpa:
  enabled: true
  minReplicas: 1
  maxReplicas: 2
  targetCPUUtilizationPercentage: 70
  scaleUpStabilizationWindowSeconds: 60
  scaleDownStabilizationWindowSeconds: 900

Key	Description	Default
enabled	Enables the HorizontalPodAutoscaler for the CodeMie API pods.	false
minReplicas	Minimum number of pods to keep running, regardless of load.	1
maxReplicas	Maximum number of pods the autoscaler is allowed to create.	2
targetCPUUtilizationPercentage	Target average CPU utilization, as a percentage of the pod's CPU request. Scaling activates when actual usage moves away from this target.	70
scaleUpStabilizationWindowSeconds	How long CPU usage must stay above the target before new pods are added. Prevents scaling up on brief spikes.	60
scaleDownStabilizationWindowSeconds	How long CPU usage must stay below the target before pods are removed. Prevents flapping when load is temporarily low.	900

info

CPU utilization is measured against the CPU request set for the CodeMie API container (see resources.requests.cpu), not the limit. Set an appropriate CPU request before enabling the HPA, since it directly determines when scaling triggers.

Pod Disruption Budget (PDB)

When enabled, the PDB tells Kubernetes how many CodeMie API pods must stay available (or how many can be unavailable) during voluntary disruptions, so a node drain or cluster upgrade doesn't take down the service entirely.

Configuration

Set the pdb block in your values-<cloud>.yaml, using either minAvailable or maxUnavailable — they are mutually exclusive:

pdb:
  enabled: true
  minAvailable: 1
  maxUnavailable: ""

Key	Description	Default
enabled	Enables the PodDisruptionBudget for the CodeMie API pods.	false
minAvailable	Minimum number of pods that must remain available during a voluntary disruption.	1
maxUnavailable	Maximum number of pods that can be unavailable during a voluntary disruption. Leave empty to use minAvailable instead.	""

warning

Set only one of minAvailable or maxUnavailable. If maxUnavailable is set, it takes precedence over minAvailable.

If maxReplicas (HPA) is set close to minAvailable (PDB), disruptions can be blocked because Kubernetes cannot evict enough pods to satisfy both constraints at once. Keep some margin between the two when running with a small number of replicas.

Applying Changes

After updating hpa and/or pdb in your values file, apply it with Helm:

helm upgrade codemie-api \
  oci://europe-west3-docker.pkg.dev/or2-msq-epmd-edp-anthos-t1iylu/helm-charts/codemie \
  --version <version> \
  -f codemie-api/values-<cloud>.yaml \
  --namespace codemie

Verify the resources were created as expected:

kubectl get hpa,pdb -n codemie