Report summarized Gitaly Apdex and Error Rate metrics via usage ping
Problem to solve
GitLab should work well for our customers in lots of different configurations. Particularly for core capabilities, like Git repositories, we want to know basic performance and availability metrics for key components so that we can identify regressions, and develop an understanding of our self-hosted customers experience of GitLab.
Further details
Proposal
The GitLab application should automatically scrape Prometheus metrics from key services at a regular frequency and provide a summary for the past 7 days in each usage ping.
Links / references
Edited by James Ramsay (ex-GitLab)