# camel-container
Docker container lifecycle component for rust-camel. Producer operations: list, run, start, stop, remove, exec, and network management (create/connect/disconnect/list/remove). Consumer mode: subscribes to Docker daemon events or container logs and emits one Exchange per event/log line.
## Language
**ContainerEndpoint**:
Endpoint for `container:<operation>` URIs; routes the operation through the Docker daemon.
_Avoid_: docker address, container address
**ContainerProducer**:
Producer that performs container operations (list/run/start/stop/remove/exec/network-create/network-connect/network-disconnect/network-list/network-remove).
_Avoid_: container operator, docker sender
**ContainerConsumer**:
Consumer that subscribes to Docker daemon events and emits one Exchange per event.
_Avoid_: event subscriber, docker listener
**ContainerTracker**:
Internal state that records created containers for cleanup on shutdown; safe for hot-reload.
_Avoid_: container registry, container state
**Auto-pull**:
When a `run` operation references an image not present locally, the producer pulls it before creating the container.
_Avoid_: image fetch, automatic pull
## Log-level policy
Per ADR-0012.
**Labels wired in Phase B (commit 5dcb80e1):**
All 4 sites are category (e) outside-contract: the Docker daemon operation failed and the route cannot meaningfully retry (daemon down, image not found, container ID invalid). Each site calls `runtime.metrics().increment_errors(route_id, label)` then logs at `error!` with `// log-policy: outside-contract`:
- `e:container:events-connect` (`src/lib.rs:1433` events consumer initial connection failure)
- `e:container:events-stream` (`src/lib.rs:1463` events stream broken mid-flight)
- `e:container:logs-connect` (`src/lib.rs:1520` logs consumer initial connection failure)
- `e:container:logs-stream` (`src/lib.rs:1602` logs stream broken mid-flight)
## Example dialogue
> "What happens if the Docker daemon goes down while consuming events?"
> "The container consumer's reconnect loop retries with exponential backoff. If all retries are exhausted, it records an error via `increment_errors` with label `e:container:events-connect`, logs at `error!` with `// log-policy: outside-contract`, and returns the error."
> "Can I track created containers for cleanup?"
> "Yes — the ContainerTracker records every container created by a `run` operation. On shutdown, it removes orphaned containers. The tracker is safe for hot-reload: stopping and restarting a route does not lose tracked containers."