steer docs

Operations

Day-two work — status, scale, resize, logs, events and images, from the CLI and the TUI.

Status

steer service status        # alias: steer service ls
steer service status -w     # refresh in place (default every 15s)

One table: desired/running/pending counts, service state, the image tag actually running and the task’s CPU/memory. Red running count = below desired; yellow pending = instances starting. The TUI sidebar shows the same, refreshed automatically.

Scale

steer service scale -s api -c 3

Sets the desired task count (confirmation included; -y to skip). TUI: s or the [ Scale (s) ] button.

Resize (CPU / memory)

steer service resize -s api --cpu 0.5 --memory 2GB

Registers a new revision with the new resources and rolls it out (-w to watch). CPU accepts 0.5, 1 or 500m; memory accepts 2GB, 2048 or 512MB. Only combos your provider actually supports are accepted — get it wrong and the error lists the valid tiers. In the TUI, z opens a picker that only shows valid combos. rollback reverts resources too.

Logs

steer service logs -s api          # last 100 lines from the last hour
steer service logs -s api -f      # keep following (poll every 3s, Ctrl+C to stop)
steer service logs -s api -n 25   # fewer lines

Zero configuration: steer discovers where the logs live from the service’s task definition (the awslogs driver). Tasks with several containers (app + sidecars) are merged into one chronological stream, each line prefixed with [container]. In the TUI, the Logs tab tails and follows the selected service automatically — and it won’t steal your scroll position while you read history.

Logs are read through CloudWatch Logs, which can only scan forward — “the last 100 lines” means the last 100 within the past hour. A service using a different log driver (firelens, splunk…) shows a clear message instead of an error dump.

Events

steer service events -s api

The service’s recent history as ECS tells it — tasks started, health checks failed, steady state reached. Last 20 events, oldest first, errors in red. In the TUI the Events tab shows them for the selected service and refreshes while you watch; during a deploy the live rollout feed takes over the tab.

Images

steer image ls              # repos: name · latest tag · pushed  (alias: img)
steer image tags -r api     # tags of one repo: TAG · AGE · SIZE · DIGEST · ● now

● now marks the tag currently running in the sibling service. Only real, deployable images are listed — dangling manifests, attestations and signatures are filtered out. In the TUI, the IMAGES section lists repos; selecting one shows its tags in the panel.