Building a Metrics Optimised Pipeline (Full Automation)
summary
Argues that continuous delivery is about idea-to-user not commit-to-production and shows how to auto-check architectural signals (SQL counts, exceptions, CPU/memory, network calls) inside pipelines so functional green tests don’t mask regressions. Describes duct-taping New Relic + CircleCI, Dynatrace, Nagios and bespoke instrumentation to baseline metrics and fail on anomalies. Gives concrete failures: a Hibernate change creating N+1 queries that multiplies DB load and a microservice refactor that increases network traffic 10x. Recommends auto-baselining via architectural metrics.
AI-written summary · about
chapters (34)
- 0:00Intro
- 1:13About me
- 1:51Whats the plan
- 3:27Audience questions
- 5:50Moving away from big bang releases
- 6:15Reducing risk
- 7:39Motivation numbers
- 8:13It doesnt matter
- 9:06Version 2 of CD
- 9:53Typical CD Pipeline
- 10:37What is a Pipeline
- 11:37Flow of Ideas
- 12:47Pipeline Structure
- 13:57Dark Launches
- 17:19Small Tweaks
- 18:07Big Leap Forward
- 18:41Multiple Pipeline Patterns
- 19:29The Point of a Pipeline
- 20:25Confidence to Release
- 21:42Confidence to Get Better
- 22:25Architectural Metrics
- 23:27Use Case
- 25:35Microservices
- 26:40Microservers
- 28:34Metrics
- 29:00Pipelines
- 29:32Extending the Pipeline
- 30:25Understanding the User Experience
- 31:05More UserFocused Teams
- 31:51Microsoft Study
- 32:15Delivering Features in Software
- 33:01Takeaways
- 33:25Zero Character
- 33:50Emojis
related talks
01h 00mDeep Dive: Self-driving Pipelines: Building the Unbreakable Continuous...
06mLightning Talk: Performance Metrics Driven CI/CD in a Nutshell
29mMeasuring Devops Success with Pipeline Analytics
55mPatterns & Anti Patterns for Scaling & Enabling DevOps Adoption
33mMeasuring Developer Efficiency in DevOps: Key Questions, Metrics, Best Practices
34mWhy continuous testing is essential
37m#DataDrivenDevops
47mArchitecting for DevOps and Continuous Delivery