Bugfix release. Surfaced by @LeeNX in #28 during his Podman test on x86-64 with podman-compose.
Fixed
Container health no longer shows as just 🟢 vs ⚪ wrongly under Podman.
/status (Telegram) and the Web UI container list both derived health icons by grepping (healthy) / (unhealthy) / (health: starting) substrings out of docker ps's Status column. That worked under Docker because the CLI appends those markers cosmetically. Podman's REST API does not — the Status field is a Docker-CLI cosmetic, not part of the OCI/Docker REST API contract. So every healthy container under Podman fell through to the generic ⚪ running icon.
Both /status (Telegram) and _get_containers() (Web UI) now batch-inspect running containers and read State.Health.Status directly — consistently provided by both Docker and Podman APIs.
Side benefit: uptime in /status is now computed from State.StartedAt instead of parsed from a Status string, so the format matches the per-container /status <name> detail view.
Cost
One additional docker inspect batch call per /status invocation (~30 inspects for a typical host, all in-memory daemon state, instant). No user-visible latency change.
Upgrade
docker pull amayer1983/docksentry:latest
docker compose up -d