How a two-week outage that only the invoice noticed changed how ClearMyInbox ships
ClearMyInbox scans a Gmail or Outlook inbox, unsubscribes from senders in bulk, then checks which ones actually stopped. The bootstrapped UK SaaS once ran for about two weeks with its API down and not a single alert firing. External monitoring on Better Stack went live in under an hour, and detection now takes minutes.
- Industry
- Email productivity SaaS
- Products used
- Uptime, status pages
- Before Better Stack
- No monitoring
The outage that only the invoice noticed
In June, before any monitoring existed, the ClearMyInbox API went down and stayed down for about two weeks. The container crash-looped the entire time: health probes failed, the app restarted, then failed again.
Nothing alerted, because nothing was watching from outside. Every crash-loop restart woke the serverless SQL database, so it never auto-paused, and that month's database bill roughly doubled. An Azure invoice was the first and only warning of a two-week outage, and it cost about £100 of pure waste.
“There is no ops team here. Better Stack is effectively it. Set up external monitoring before your first user, right after you buy the domain.”
Monitoring, same day
As a solo founder, Chris is the entire engineering and ops team, and external monitoring had never reached the top of the list. Once it became urgent, he looked at what other indie SaaS products use and kept seeing Better Stack status pages under bootstrapped products. There was no evaluation process, because he needed monitoring that day.
Setup took under an hour: three uptime monitors covering the site, the API and a second health check on the scheduler endpoint that drives the weekly scans, then a public status page on a custom domain at status.clearmyinbox.ai.
The only ops report worth reading
The weekly summary email has quietly become the whole of ClearMyInbox's ops reporting. All three monitors have recorded 100% uptime since setup, and that silence now carries information it never could before.
“The weekly no-downtime-this-week email is my favourite kind of boring.”
What changed
External uptime monitoring now covers the site, the API and the scheduler endpoint, independent of the cloud provider that never alerted. The status page is branded and public. Detection takes minutes rather than weeks, and the hours a week that went into manually checking dashboards go into building the product instead.
“I ship knowing that if the API stops answering, I find out before my customers do.”
Uptime monitoring, incident alerting and a branded status page in one platform.