Why our status page is now just a prayer
By Jonas K., Head of Decoherence
There is a well-known paradox in distributed systems: the act of observing service health degrades it. Every uptime check is a small prod, a tiny perturbation. At sufficient scale, the cumulative effect of monitoring is itself the largest source of downtime.
After studying our 2025 incident reports, we realized that 71% of our outages began within nine seconds of a customer refreshing status.splitthebit.com. The status page wasn't reflecting outages. It was causing them.
Our solution: we replaced the status page with a single rotating quote from Marcus Aurelius and a candle.
Since the launch of status.splitthebit.com/v2, customer-reported incidents are down 64%. (Internal incidents are up 220%, but those don't count.) We consider this a clear win for both our infrastructure and the broader engineering discipline.
Disclaimer
This post may have been written. The author may exist. Forward-looking statements herein are not guarantees of any specific timeline.