Metrics Driven Blue-green Deployments using Spinnaker’s Cloud...
summary
Spinnaker’s Cloud Foundry integration deploys a baseline and a canary, uses Kayenta canary analysis against DataDog metrics (prediction_time.average) and blocks promotion on regressions. Demo pipeline: Jenkins builds to GCS, Google Pub/Sub triggers Spinnaker, metric tags distinguish production/baseline/canary, micrometer emits timings. Requirement is ~200ms; a bad model spikes requests from ~250ms baseline to ~650ms and fails the canary; after fixing the delay the canary matches baseline (1.4% deviation) and Spinnaker promotes and cleans up.
AI-written summary · about
chapters (13)
- 0:00Introduction and setup
- 1:07What is Spinnaker
- 2:56Spinnaker Cloud Foundry integration
- 6:28Resource mapping in Spinnaker
- 8:38Live demo of Spinnaker UI
- 13:20Building a deployment pipeline
- 15:24Introducing the demo app
- 17:34Pipeline trigger and flow
- 23:13Deploying baseline and canary
- 27:30Manual judgment and canary analysis
- 31:25Canary analysis results
- 32:52Deploying to production
- 35:58Summary and Q&A
related talks
35mCutting-edge Continuous Delivery: Automated Canary Analysis on CF using Spinnaker
51mCutting edge Continuous Delivery: Automated Canary Analysis through Spring based Spinnaker
52mContinuous Deployment to the Cloud using Spinnaker
35mReduce Risks With Continuous...
12mImplementing Advanced Deployment Using Spinnaker and Canary
36mSpinnaker - Land of a 1000 Builds
21mI Don’t Always Test My Code, but When I Do It’s in Production
24mUsing Jenkins and Spinnaker to Supercharge Your Deployment Pipelines