The pattern was tab fatigue
Over the holidays we read 100 publicly written incident reviews — Cloudflare, GitHub, Shopify, Discord, smaller startups. The pattern that surprised us wasn't query latency. It was tab fatigue.
In 73 of 100 reviews, the responder mentioned switching between three or more dashboards or query interfaces during the active incident. In 41, they specifically called out "we lost time pasting trace IDs / timestamps" or equivalent.
This is the gap molesignal exists to close.
A few other findings, sorted by surprise
- 18 reviews praised their tooling's UI explicitly. The rest either complained or didn't mention it.
- 9 reviews said the observability cost was a factor in choosing not to retain enough data to debug fully.
- 6 mentioned that bringing context to on-call was the most expensive part of being on-call.
- 0 reviews praised the alert evaluation latency. (Conspicuous silence.)
We're publishing the open list of links in /docs after v1.0. Until then, if you've read more than we have, come tell us at /design-partner.