Features

Feature	Description
Auto-update	Polls a local Docker registry; redeploys a compose service the moment a new image digest lands
Rollback	Tags the previous image before every deploy; reverts automatically if the new container fails to become healthy
Auto-heal	Listens to Docker health events; restarts unhealthy containers with per-service cooldown to prevent restart storms
Multi-image services	One service entry can track multiple images; a single compose redeploy fires when any image changes
Heal-only mode	Watch and recover a service without ever touching its image — useful for third-party or pinned containers
Blocked digest protection	Blocks a known-bad digest in memory; dockward skips it until a new digest is pushed to the registry, the service is manually unblocked via API, or dockward restarts
Resource alerts	Per-container CPU and memory thresholds with independent cooldowns; fires notifications before a service falls over
Audit log	Append-only JSONL file of every deploy, rollback, restart, and alert — full trail for post-mortems
Live web UI	Real-time dashboard over SSE; shows service health, deployed image, container list, and recent events without page reloads
Prometheus metrics	Counters and gauges for updates, rollbacks, restarts, and failures; drop into any existing Grafana setup
Notifications	Discord webhook, SMTP email, and arbitrary HTTP webhooks; each event type is independently configurable
HTTP API	`/status`, `/trigger`, `/unblock`, `/audit`, `/metrics` — automate or integrate with any toolchain
Warden mode	Central aggregator for multi-machine fleets; agents push audit entries over HTTP, warden exposes a unified SSE dashboard
Single binary	Zero runtime dependencies, ~12 MB; ships as `linux/amd64`, `linux/arm64`, and `darwin/arm64`
Compose-native	Delegates all deploys to `docker compose up -d` — no custom orchestration, no vendor lock-in
Systemd-ready	Ships with service units for both agent and warden modes; `systemctl enable dockward` and done