System status
All systems operational
Checked live at Fri, 21 Aug 2026 18:03:33 GMT
Website & app
Serves the public site and the signed-in web app
Database
Stores your automations, run history, and approvals
Automation scheduler
Runs your scheduled and event-triggered work
Task execution health
Detects automation runs that hang or crash mid-run
99.95% uptime
Over the last 90 days, from 2,007 recorded checks. Counted from the heartbeat only — it runs on an even cadence, so every sample represents the same amount of time. This figure counts whether Sprawl was serving. It does not count a scheduled run of yours arriving late, which is reported separately below.
89.14% of those checks had every internal system green — the stricter number, including background jobs like the scheduler and our queue provider. It runs lower than uptime on purpose, and some of the gap is automations waiting on an integration a customer disconnected rather than anything being broken.
99.60% of the checks that should have run, did — 2,007 recorded against 2,015 expected at the heartbeat’s 30-minute cadence. This is the number that matters most and the one status pages usually omit: the checks are written by a cron running inside Sprawl, so a total outage records no check at all rather than a failed one. It disappears from the two percentages above and shows up only here.
When one of your runs fails
The number above is about whether Sprawl was serving. This is the other half: what you are told when your own automation goes wrong. Both are reliability, and only one of them shows up on a status page anywhere else.
A failure says whose it is.
If the tool that broke is one of ours, the run says so rather than sending you off to reconnect an integration that was working: A Sprawl tool is broken (…) — nothing to reconnect, it is ours to fix.
A partly-failed run is counted, not rounded.
A run where some steps worked and others did not is not flattened to success or failure. It reads like Partly failed — 2 of 5 steps failed: Gmail (gmail_send_email). Sprawl read this from the run itself. — with the count, the tools named, and an explicit statement that Sprawl measured it rather than the agent claiming it.
The honest limit: both of these live inside the product, so you cannot check them from here without an account. They are described on this page precisely so the claim is on the public record before you sign up, and can be held against us afterwards.
These checks run against live infrastructure every time this page loads — nothing is cached or hand-set.