Argus watches capacity across every region of your cloud on a fixed interval and renders it as one living heatmap — so you see, at a glance, where your instances are most likely to launch and least likely to be reclaimed.
In myth, Argus Panoptes never slept — a hundred eyes, forever watching. This one keeps watch over your capacity. PANOPTES · “THE ALL-SEEING”
A single self-hosted Go binary. Point it at your cloud account, list the instance types you care about, and let it watch.
On every tick, Argus checks each instance type across your regions and records how likely it is to launch — batched into as few requests as possible.
Scores land in SQLite. A failed fetch never wipes good data — the last known score stays put, its original fetched_at intact.
One heatmap. Red means an instance is unlikely to launch, green means it almost certainly will. Hover any cell to see exactly when that score was last observed.
No dashboards to wire up, no time-series database to run. Just scores, kept fresh and rendered honestly.
Instances × regions, colour-graded 1–10, updating on its own interval.
Score any instance type across all regions enabled for your account, or a curated subset.
Throttled or hit a transient error? Existing scores survive, untouched — never blanked out.
Durable, queryable, zero-ops storage that survives restarts. No external database required.
One Go executable with the UI embedded. Self-host it anywhere; ships as a lean Docker image too.
A /health endpoint out of the box, and a schema built to span more than one cloud.
Mount a config, hand it your cloud credentials, and open the dashboard on :1342. That's the whole ritual.
# watch your capacity in one command docker run -p 1342:1342 \ -v $(pwd)/config.yaml:/app/config.yaml:ro \ --env-file .app.env \ ghcr.io/codingcoffee/argus # → dashboard at http://localhost:1342/dashboard
Let a hundred eyes watch your capacity for you — continuously, honestly, in one glance.