getsentry
Make it possible to extract and analyze performance data - This is possible with errors today via webhooks
Problem Statement
LaunchDarkly recently created an integration with Sentry to automatically ingest error events from the platform, and attribute those errors with LD contexts for use in the LD experimentation platform.
One thing we were additionally interested in adding was performance metrics (latency, in particular). This was said to not be possible at the time of integration, and so we only integrated with errors.
It would be great if we could eventually handle performance data in the same way and provide LD customers with sentry latency events attributed to flag evaluations.
Solution Brainstorm
Surface latency webhook for integration use as a solution?
Product Area
Settings - Integrations
┆Issue is synchronized with this Jira Improvement by Unito
![getsantry[bot] avatar](https://avatars.githubusercontent.com/in/66573?v=4 "Published by a pub.dev verified publisher"){kind=small}
Hi @cdelst Thanks a lot for bringing up this use case- As you mentioned we have error webhooks but nothing for our performance data today. I will add this to our backlog and keep this issue open to hear from the community
A lot has changed since Jul 2023. Launnchdarkly has a new feature called Guarded Rollouts which couples very nicely with perf metrics from sentry.
What this feature does is progressively roll out a flag to a percentage of users and looks at metrics you specify to detect regressions. The metrics could be generated with their track api but they also support sending your open telemetry spans to them and they auto generate metrics from it.
A really nice feature, if sentry and launchdarkly could coordinate it, is ingesting sentry's otel spans and auto generate metrics in LD for them. This would enable auto rollback of a feature flag that degrades web vital or perf metrics 😍
