Skip to main content

Container Stack Deployment

Use this tutorial when you want an agent to perform a stack rollout with explicit checkpoints.

Copy-paste prompt​

Deploy or update my container stack on mittwald and validate runtime health.

Flow:
1) Inspect existing stacks, containers, and volumes.
2) Deploy/update the target stack.
3) Verify service health, ports, and persistent volumes.
4) Report rollout status and any failing service IDs.

Do not scale or delete resources unless I approve.

What the agent will do automatically​

  • Inventory existing stack/container resources.
  • Execute deployment/update steps.
  • Validate runtime and persistence status.
  • Produce a rollout summary with IDs.

What you (human) must still do​

  • Approve scaling/deletion changes.
  • Provide missing env or secret values.
  • Decide rollback if health checks fail.

Likely questions the agent will ask and good answers​

  • Which stack should I deploy? -> Use stack <stack-name> in project <project-id>.
  • Can I scale services? -> Scale only after initial health is green.
  • Should I prune unused resources? -> No, report only.

Verification prompt​

List stacks, containers, and volumes for this project and confirm health status after deployment.
Return IDs and failing checks, if any.

Rollback/cleanup prompt​

Rollback the stack to the previous known-good configuration and confirm container health after rollback.

Resume after failure​

Resume the deployment from the failed stack step only. Keep already successful validation steps as complete.