What each check does
This page runs on its own servers, apart from the app. Every minute it checks Freesbe fromFrankfurt (Europe), Virginia (United States), Singapore (Asia-Pacific) and Sydney (Australia), the way your app does it: it signs in, sends a message, uploads a file. Nothing on this page waits for someone at Freesbe to notice a problem and post it.
The parts
| Part | Every minute, from every place |
|---|---|
| Web and desktop app | app.freesbe.net serves the app, and a test account signs in through the API. |
| Messaging | A message is sent in a test workspace stored in that region and must arrive over the live connection within five seconds. Then it's deleted. |
| Files and uploads | A small file is uploaded to the region's test workspace and downloaded again. |
| Calls and meetings | Every call server answers its health check, and the call relay hands out credentials. No call is placed. |
| Phone system | The phone network's API answers, and the share of calls the network reported as failed in the last hour. Nothing is dialled or bought. |
| AI features | The share of AI requests that failed in the last 15 minutes, counted by the app. |
| Email and notifications | The share of emails and push notifications delivered in the last hour, counted by the app. |
| API and integrations | api.freesbe.net answers its health check, and the share of webhooks delivered in the last hour. |
Working, degraded, down
- Working: the last check passed everywhere, and nothing is open on it.
- Degraded: the last check failed in one place, or more than 2% of requests failed. One slow minute shows here first.
- Down: two checks in a row failed, or more than 10% of requests failed.
- Maintenance: planned work our team posted ahead of time is under way.
- No data: this page couldn't get a reading, for example when the app can't be reached to count its requests. It never shows green for a part it hasn't measured.
Our team posts incidents on top of the checks, with updates as they go and what happened afterwards. An incident can mark a part as degraded or down even when the checks pass, for trouble the checks can't see.
The 90-day bars
Each bar is one day, in UTC: green when every check that day passed, amber when some failed (or an incident touched it), red for a day with an outage, blue for planned maintenance, grey when nothing was checked. For AI features and email, which are measured by counting rather than by a check, each minute the counts were read is one check, failed when the rate was over 2%. The uptime figure is the share of checks that passed over the 90 days, every place together. On a phone the bars show the last 30 days.
What our infrastructure provider reports
Freesbe runs on a large infrastructure provider that publishes a status page of its own. When it reports trouble or planned work on a service Freesbe uses, the parts of Freesbe it touches show a short note. The note is context: this page's own checks decide the colour.
When this page is down too
Then write to support@freesbe.net. Machines can also read/api/summary.json, which answers whenever this page does.