Status

This page reports two different things and does not mix them up. The first is the handful of services we run ourselves: the website you are reading, the shared PostgreSQL behind it, the killfeed ingest that streams killmails into our own archive, the job queue that runs translations and syncs and purges, and the Discord bot that posts announcements on our servers. When one of those is unwell, it is our problem and the headline at the top says so.

The second is everything we depend on and cannot fix: CCP's ESI API and the EVE Online SSO login, eve-kill for killboards and pilot statistics, EVErefs for the reference prices behind every ISK figure, EVE Scout for Thera and Turnur, Discord, Cloudflare R2 for images, Twitch, YouTube, EVE Workbench and Reddit. Those never colour the headline — a Reddit outage is not a Capsuleers.app outage, and the page that uses it degrades gracefully on its own — but each one carries its own state, so you can tell whether the gap you are looking at is ours or theirs.

Nothing here is checked because you opened the page. Probes run on their own schedule and you read the last result with the time it was taken, because a status page that queried fifteen services on every visit would become part of the outage it is meant to report. Where a scheduled job of ours already uses a service, we do not probe it at all: the queue records whether the last run succeeded, which is better evidence than a ping and costs nobody a request. Every row says which of the two it is.