Operations

Architecture choicesWhich runtime to run where, and the one configuration that must not be duplicated.

Scheduled executionThe CI cron that does the actual monitoring.

API hostingContainer hosts, cold starts, and the alternatives.

Panel hostingDeploying the static admin panel.

MonitoringHealth alerts, run history, and what silence means.

Backup and restoreThe database is the only thing you cannot rebuild.

Costs and quotasWhat runs out first, and what happens when it does.

SecurityAuth model, secret handling, exposure surface.

UpgradingUpdating an instance without losing state.

The operating principle

The monitoring runs as a short-lived outbound job; the API and panel are optional conveniences. If you remember one thing, remember that: if the API is down, your watches still run. That is what keeps the whole system cheap and hard to break.