Don't GitOps Into a Blackhole
Talkat GitOpsCon North America 2024 · on Monday, 2024-04-15· at 15:20· Level 3 | Room 322 | Track 2·
summary
Keptn proves a concrete GitOps extension: CRDs define pre-deploy CPU >1 check and post-deploy tasks so deployments block or warn based on live metrics. A demo reads available CPUs from Prometheus (cluster shows 2, check passes), runs a post-deploy console task, and succeeds. Keptn’s metrics operator averages a 5‑minute CPU‑throttling query (returned ~2.83) and metrics operator exposing values to HPA causes replicas to scale from 1 to 3. Also integrates with Argo rollouts and other providers.
AI-written summary · about
chapters (7)
- 0:00GitOps risks
- 1:45Speaker and Kepon project
- 3:10Kepon lifecycle and metrics
- 5:23Pre-deployment task evaluations
- 11:25Observability challenges
- 12:48Metrics integration with HPA
- 18:12Future Kepon capabilities
related talks
19mWorking Your Way Through GitOps with Keptn
24mOrchestrating and Observing GitOps Deployments with Keptn
29mObservability Driven Automation: Beyond GitOps with Keptn
36mKeptn Beyond 1.0: Sailing into the Future
36mA Guide to Event-driven SRE-inspired DevOps
18mPerformance Testing in Keptn Using K6
23mGitOps: yea or nay?
27mPutting Chaos Into Continuous Delivery to Increase App...