Resource Limits

For larger deployments it is sometimes hard to keep track of the defined resource limits. For example a deployment for a distributed system having lots of similar workloads are suitable for a global definition of resource limits. adeploy does support such a global definition at a single point in your default or namespace/release configuration.

Global Resources Configuration

Jinja

adeploy introduces the reserved variable __resources that allows to define resource limits for all workloads on a global, workload or container scope.

Global Scope

To set resource limits for all compatible API objects (Deployment, StatefulSet, ReplicaSet, ...) in your deployment, add the following to the defaults.yml or the namespace/release configuration:

defaults.yml

_resources:
  limits:
    cpu: 1000m
    memory: 265Mi
  requests:
    cpu: 500m
    memory: 128Mi

Note

These globally set limits are ignored, if limits are already defined in the manifest files for the workload objects in the templates folder.

Workload Scope

To set resource limits for a specific API object queried by its name (i.e. nginx), add the following:

defaults.yml

_resources:
  nginx:
    limits:
      cpu: 1000m
      memory: 265Mi
    requests:
      cpu: 500m
      memory: 128Mi

Note

These globally set limits are ignored, if limits are already defined in the manifest file for the workload object nginx in the templates folder.

Container Scope

To set resource limits for a specific container queried by its name (i.e. main) and the workload object (i.e. redis), add the following:

defaults.yml

_resources:
  redis:
    main:
      limits:
        cpu: 1000m
        memory: 265Mi
      requests:
        cpu: 500m
        memory: 128Mi

Warning

These globally set limits take precedence and will overwrite any limit that is defined in the manifest file for the image main of the workload object redis.