Zero Downtime Application Deployments on Azure PaaS
summary
Shows how to get true zero downtime on Azure PaaS by using Azure DevOps pipelines, Azure Web Apps deployment slots (up to 20, cookie‑based routing and query‑string pinning) and AKS rolling updates with readiness probes and MaxUnavailable=0. Adds observability via HealthChecks, OpenTelemetry, Prometheus/Grafana or Application Insights to detect latency/CPU spikes, uses feature flags to separate deploy from release and a DB playbook: deploy code that supports both schemas, add nullable columns as zero‑op, migrate gradually and remove old schema. Advocates releasing ~30 times/day.
AI-written summary · about
chapters (10)
- 0:00Zero downtime deployments
- 1:07Key ingredients
- 13:44Azure Web Apps slots demo
- 19:46Kubernetes fundamentals
- 35:41Adding observability
- 43:33Kubernetes rolling updates demo
- 51:01Feature toggles
- 55:14Deployment strategies
- 59:26Database schema changes
- 1:03:18Conclusion and key takeaways
related talks
40mStrategies for Zero Down Time, Frequent Deployments
46mReal World DevOps
01h 00mEasier Deployments using a Green:Blue Strategy
01h 01mFrom one release per quarter to 30 times a day
37mZero-downtime Deployment Strategies (...) [ENG]
53mContinuous Delivery with Azure Web Apps
45mRunning a Real-World, Mission-Critical System on Azure
32mZero Downtime Deploys Made Easy