Wow, about an hour ago we started getting flakey alerts that our APIs were down. We're running on Lambda so I jumped in and started searching (average duration was well over 5s, usually <50ms). Opened HN to see if Lambda was having issues, saw this and realised that we use Sentry's APM, and it has been causing our application requests to take >10s!
Ironic that the performance monitoring tool obliterated our performance.
https://cln.sh/dUMc05odMEREo9HC9E0u