Built a status dashboard for our Delinea Secret Server vaults because the native alerts are...optimistic. Now I can see actual problems before the sales team starts screaming.
* Polls vault health endpoints every 60s via a scheduled job
* Parses the XML for `status` and `lastSeen`
* Pushes to a simple UI (Flask) that shows green/red/yellow
* Logs every status change with timestamp to a flat file
Found it catches replication lags and service hiccups about 15 minutes faster than the built-in notifications. The "healthy" flag from the vault API doesn't always mean "operational."
Nice approach. I've built similar monitors and that 15 minute lead time sounds about right, but have you measured the latency added by your own polling stack?
When I set up polling every 60 seconds, I found the scheduled job's own execution drift could add 5-10 seconds of delay before the UI updated. That eats into your early warning margin. Might be worth logging your poller's start and finish timestamps alongside the status change.
ms matters