@strangeloop_stl@frankc love this story: added traces on the 2nd day of a 2-day incident, and found the critical clue within 2 hours. ("what's different between the weirdly behaving parts of the fleet... and everything else?" ✨) (@frankc#strangeloop)
@strangeloop_stl@frankc "we were able to look at telemetry with context." - and the ability to separate out thousands of test runs, and thousands of PRs, to pinpoint the effect of a single, small change (@frankc#strangeloop)
• • •
Missing some Tweet in this thread? You can try to
force a refresh
going through painful set of High Crimes of observability with @adrianamvillela at #monitorama
v on board with this 🙌
traces are a building block *for* metrics: more flexibly and powerfully than relying on pre-aggregated metrics themselves
🙈 base your workflows around SLOs (the things you’ve already agreed are important signals avoid your biz/org!), not on artifacts that you captured the last time something went wrong