Current state

All systems operational

Events are accepted at the ingest edge, stored before any delivery is attempted, and delivered to the destinations you configure.

Webhook ingest, last 90 days: calculating
All components, last 90 days: calculating
Method: Checks passed over checks run.

Operational

Degraded, answered but not healthy

Not responding

Not recorded

Each row's state comes from our recorder's latest check, which runs from one region every fifteen minutes, as do the ticks, uptime figures, and round trip line. The probe line is measured from your browser and includes your own network. The delivery pipeline has no public endpoint of its own; the API answering stands in for it, and the row says so rather than claiming a measurement that is not being made.

Run the same checks

Every row above is a request you can make yourself. These print the status code and the round trip.

Webhook ingest

curl -sS -w '\n%{http_code} in %{time_total}s\n' https://hooks.shellorbit.com/health

Dashboard API

curl -sS -w '\n%{http_code} in %{time_total}s\n' https://api.shellorbit.com/health

Send a test event to your own endpoint
curl -sS -X POST https://hooks.shellorbit.com/e/$INGEST_KEY \
  -H 'content-type: application/json' \
  -d '{"ping":true}'

This page, for a script

curl -sS https://status.shellorbit.com/index.json

Incidents

No incidents recorded.

Anything affecting receipt or delivery is posted here, with what happened and what changed afterwards.

What is measured

Availability: The proportion of well formed requests to the ingest endpoint that receive a 2xx acknowledgement. Receipt is what the product promises: an acknowledged request is stored, and a destination being down never puts it at risk.
Current state: The headline and each row come from the recorder's latest check, which is never more than about fifteen minutes old; the line above the rows says how old. A failure your browser can read for itself, from the ingest endpoint, is newer and takes precedence. If the recorder has not reported for 45 minutes, the page stops treating its last answer as current and falls back to your browser's probes.
A day is one tick: Our recorder checks every component every fifteen minutes, ninety six times a day, and each tick carries how many of those passed. Green at 98 percent and above, which allows one or two missed checks in a day; amber below 98, meaning something was wrong for more than half an hour; red below 90, meaning most of the day was lost. Hover a tick, or focus the strip and use the arrow keys, for the count. Unfilled means we were not recording, never good news.
Uptime figure: Checks passed over checks run, across the recorded window, for the ingest endpoint on its own and for every component together. The recorder runs from a single region, so a figure here is a lower bound on what a request from that region saw and not a promise about yours.
Commitment: Every plan carries a 99.9 percent monthly availability commitment on the ingest endpoint, the free plan included. Paid plans are credited for a shortfall. Exclusions are set out in the terms of service.
Not counted: A destination of yours refusing a delivery is not an outage here. Those events are stored, retried, and visible in your event log. Delivery failures belong to the destination, and the dashboard shows them per endpoint.
Machine readable: index.json lists every probe and what a healthy answer looks like, alongside the incident history. uptime.json is the recorder's own file: the daily counts, the time of the latest check in updated_at, and each component's latest verdict in latest.