Synthetic checks, alerting, and nothing you did not ask for.
The service exists because monitoring hosted inside the same infrastructure as the monitored system tells you nothing during the outages that matter most. Our probes sit with independent providers across several countries, and the control plane is deliberately boring.
SMS dispatch queued for up to 11 minutes after an upstream aggregator throttled our sender. Webhook and email delivery were unaffected. A second aggregator has been added and SMS now fails over automatically.
A routing change at the Dublin provider caused intermittent timeouts to a subset of destinations. The confirmation layer suppressed most alerts as designed, but 14 checks were briefly reported as failing. probe-ie...
Planned migration of the history store. The API returned 503 for a 20 minute window during cutover. Check execution and alerting continued throughout.