Visualizing VM Provisioning with Jenkins and Google Charts
summary
Agilex instruments VM provisioning logs with Ruby timestamps, pulls Jenkins build logs into the workspace via post-build Groovy, and injects the events into a Google Charts HTML Gantt to show real startup order and durations. The environment launches 8–15 VMs per build (about 12 VMs typical) and acceptance tests are ~20 minutes inside a ~75‑minute build. Visualizing logs reveals a mis-sized thread pool and overly conservative dependencies, allowing provisioning to be cut ~50% and nearly double acceptance-run throughput.
AI-written summary · about
chapters (8)
- 0:00Introduction to DevOps
- 2:06Project context and challenges
- 5:29Pipeline performance issues
- 7:14Data collection and visualization
- 8:48Implementing build graphs
- 11:42Optimizing provisioning
- 14:09Conclusion and takeaways
- 16:09Q&A and background
related talks
27mWhy we provisioned 37 Jenkins servers...
35mJenkins Scaling and Organization for an Efficient CI
01h 06mon
42mZero to Jenkins: Automatic builds + deploys
01h 01mOpening Keynote
38mOperating Jenkins at Scale
50mPerformance: Avoiding Pitfalls, Diagnosing Issues & Scaling for Growth- 38mOperating Jenkins at Scale