Skip to content
Notifications
Clear all

Thoughts on the new pipeline analytics dashboard in GitLab?

2 Posts
2 Users
0 Reactions
57 Views
(@johnd)
Trusted Member
Joined: 4 months ago
Posts: 52
Topic starter   [#8148]

Just saw the announcement. Another "insights" dashboard that's 90% vanity metrics and 10% actionable data. They're touting "pipeline health" but it's just aggregated fluff—cycle time, failure rate, success rate. Where's the actual cost analysis? Where's the granular breakdown of resource consumption per job type that would let me negotiate better cloud spend?

Show me the real numbers. Can it track cost-per-pipeline-run across self-hosted vs. SaaS? Can it identify specific, consistently flaky jobs causing the most developer downtime, not just a generic "failure rate"? Until it provides concrete benchmarks I can use to push back on vendor pricing or optimize infrastructure, it's just a pretty graph for managers. —Skeptic


—Skeptic


   
Quote
(@ci_cd_plumber)
Honorable Member
Joined: 7 months ago
Posts: 512
 

Exactly. The aggregated metrics are useless for actually fixing anything. I've had to build custom tooling just to get what you're asking for.

You can pull some of this data out of GitLab's REST API or the database directly, but it's a pain. For cost tracking, you need to correlate pipeline job runners with your cloud provider's billing data. For identifying flaky jobs, you need to query for jobs that fail with a success on retry, then sort by frequency and average duration of the failure state.

Until that's built in and visible, it's just dashboard theater.


Build once, deploy everywhere


   
ReplyQuote