Most DevOps diagrams are too simple. The Venn diagram with something in the middle, the “wall of confusion” with dollars at the end, the myopic view of a tool sitting between dev and ops. The most useful model is the three ways: see the system, focus on flow, build feedback loops, and continuously improve. This talk introduces a practical technique for doing exactly that.
Monitoring stream mapping is a reverse value stream mapping. Start from production, not from the backlog. Get all silo groups in one room with Post-its on a big wall (not a whiteboard – you need to move things around). Map production components first. You will be surprised how many people discover components they did not know existed. Then add external dependencies (cloud services, Gmail, ticketing systems), internal dependencies (firewalls, networks), and supportive IT (monitoring systems, backup systems that nobody treats as production but absolutely are).
Then map the people: who supports what, who gets alerts, who fixes production issues, and how they interact. Cross-reference people with components. Ask which end-user functionalities cross which components – this is where the question marks appear, because many teams have never actually used their own system. Continue into the project quadrant: if something needs fixing, who do we ask? What is the development process?
The four areas of improvement that emerge: extending delivery to production (CI/CD), extending operations feedback to the project (monitoring, metrics), embedding project knowledge into operations (so ops understands business context), and embedding operational knowledge into the project (getting non-functional requirements on the backlog). Each area has three layers (tools, process, culture) and maturity levels. The bottleneck moves – fix one area and another becomes the constraint. Infrastructure as code is a conversation starter, not just an automation tool. Vagrant gives developers production-like environments. StatsD lets developers add custom metrics without asking ops. ChatOps creates shared history. Game days surface bottlenecks. And remember: the bottleneck might not be in tech at all – it might be in HR or finance.
Watch on YouTube – available on the jedi4ever channel
This summary was generated using AI based on the auto-generated transcript.