tumult-containers — Container Chaos

Script-based plugin for Docker and Podman container chaos.

Prerequisites

• docker or podman installed and accessible

Actions

Action	Description	Key Variables
kill-container	Kill a container (SIGKILL)	TUMULT_CONTAINER_ID (required), TUMULT_SIGNAL, TUMULT_RUNTIME
stop-container	Graceful stop	TUMULT_CONTAINER_ID (required), TUMULT_TIMEOUT, TUMULT_RUNTIME
pause-container	Freeze all processes	TUMULT_CONTAINER_ID (required), TUMULT_RUNTIME
unpause-container	Resume frozen container	TUMULT_CONTAINER_ID (required), TUMULT_RUNTIME

Probes

Probe	Description	Output
container-status	Container state	running, stopped, paused, not_found
container-health	Health check status	healthy, unhealthy, starting, none, not_found

Runtime Selection

Set TUMULT_RUNTIME to docker (default) or podman:

method[1]:
  - name: kill-db-container
    activity_type: action
    provider:
      type: process
      path: plugins/tumult-containers/actions/kill-container.sh
      env:
        TUMULT_CONTAINER_ID: postgres-primary
        TUMULT_RUNTIME: podman

Example Experiment

steady_state_hypothesis:
  title: API container is healthy
  probes[1]:
    - name: check-api
      activity_type: probe
      provider:
        type: process
        path: plugins/tumult-containers/probes/container-health.sh
        env:
          TUMULT_CONTAINER_ID: api-server
      tolerance:
        type: exact
        value: healthy

method[1]:
  - name: kill-api
    activity_type: action
    provider:
      type: process
      path: plugins/tumult-containers/actions/kill-container.sh
      env:
        TUMULT_CONTAINER_ID: api-server

rollbacks[1]:
  - name: restart-api
    activity_type: action
    provider:
      type: process
      path: docker
      arguments[2]: start, api-server