The speaker started by describing the daily routine of a developer working with CI/CD pipelines and the challenges they faced in identifying and resolving issues. They then introduced their company and their role as a Principal Developer Advocate. The main focus of the talk was on how they improved the observability of their CI/CD pipeline using open source tools like Elasticsearch, Kibana, OpenSearch, OpenTelemetry, and Jaeger. They explained the four steps in their process: collect, store, visualize, and report/alert. They also discussed the importance of defining key metrics and using them to guide the creation of dashboards and alerts. The speaker showed examples of different visualizations and explained how they monitored flaky tests and performance using distributed tracing. They discussed the importance of monitoring system and environment metrics to identify issues not related to the released code. The talk concluded with a Q&A session where the speaker answered questions about data persistence and surprising insights gained from implementing observability.