Is the API up?

The page fetches GET /api/v2/health and GET /api/v2/health/history, both unauthenticated, and asks again every minute. The API answers each from a short cache, so what you see can be a few seconds behind.

Database is a real query and its round trip. The other three rows are background workers, each of which writes a heartbeat at the end of every loop. A worker counts as healthy while its last finished loop is within three of its intervals, and never less than two minutes, so a queue that takes a minute to drain is not called an outage.

The API samples its own health once a minute and keeps 90 days of those samples. One bar is one UTC day: green at 99.9 percent and above, amber from there down to 90, red below 90. Hover a bar, or read it with a screen reader, for the exact number and how many minutes were degraded.

A day is measured against every minute it should have had, 1440 of them, and today against the minutes already behind it. A minute with no sample counts as down, because an outage big enough to stop the sampler is still an outage. The percentage at the top of the strip is the whole window weighted by those minutes.

Days before we started sampling are simply not there. We would rather show you a short strip that is true than a long one that is filled in.

Green means our side answered. It does not prove that delivery to LinkedIn, X, Bluesky, Discord, Telegram or Mastodon is healthy: a platform can be rejecting posts while every check here stays clean. Those failures show up per post, as a failed item with a platform error code, not here.

The fetch runs in your browser, so a red result can also be your network or an extension. If the page says it cannot reach the API, try the endpoint from a terminal before you assume it is us.

Start with the response, not this page. Every yeetpost error carries a code and, for the ones worth chasing, a req_id:

1. Read the error code against the error reference. Most failures are a bad slug, an expired connection or a platform rejection, and the code says which.
2. Check the rate limit headers. The limit is 60 requests per minute per API key, per endpoint group. See rate limits and request ids.
3. Still stuck? Send us the req_id from the response. It lets us find that exact request.