ProbeKit status
Live health of the ProbeKit website, API, and background services. Every service is checked once a minute, and the people who run ProbeKit write up every incident here.
Loading the latest checks…
How this page works
Each service is checked once a minute from inside ProbeKit. The website and API check themselves and the database, the worker records every scheduled run and whether monitor checks are running on time, DMARC ingestion reports whether the inbound mail listener and the report mailbox are healthy, and the mail relay is checked for its greeting without sending any mail.
A minute with no check counts as down for the website, API, and worker, so a stopped process shows as an outage instead of silence.
The worker also counts its real work: it shows as degraded when scheduled monitor checks run more than 15 minutes late or one check holds the worker longer than its time limit, and as down when checks are an hour late.
Uptime is the share of measured minutes a service was available, per UTC day. Degraded minutes count as available; hover or tap a day for its numbers.
When the outside check is on, a separate machine on the internet requests ProbeKit’s readiness endpoint every minute and reports what it saw, so this page does not rely only on ProbeKit reporting on itself.
Questions about an incident? Contact us.