Jenkins - Metrics and Trends

Jenkins - Monitoring Metrics & Trends

Understanding the performance and trends of your Jenkins builds and the health of the Jenkins instance itself is crucial for maintaining an efficient CI/CD platform. Jenkins offers several ways to track metrics and visualize trends.

Built-in Build Metrics

• Build History: The build history list on a job’s page shows recent build statuses (success, unstable, failure), durations, and links to console logs/artifacts.
• Test Result Trend: If using the junit step, a graph showing the trend of passing, failing, and skipped tests over time is displayed on the job page and test result page.
• Code Coverage Trend: Code coverage plugins (JaCoCo, Cobertura via Coverage API) often display trend graphs showing coverage percentages over recent builds.
• Static Analysis Trend: The ‘Warnings Next Generation’ plugin shows trend graphs for the number of issues detected by analysis tools.

Pipeline Performance Visualization

• Pipeline Stage View (Classic UI): The default view for Pipeline jobs shows the duration of each stage, helping identify bottlenecks within your workflow.
• Blue Ocean Plugin: Provides a more sophisticated graphical view of pipeline execution, including stage durations, parallel step timings, and easy access to logs per stage.
• Pipeline: Stage Step Measures Plugin: Adds metrics about stage durations to the classic Stage View.

Job-Level Metrics Plugins

Build History Metrics Plugin

This plugin calculates and displays simple reliability metrics directly on the job page:

• Mean Time To Failure (MTTF): Average time builds stay successful before failing.
• Mean Time To Recovery (MTTR): Average time it takes for a job to become successful after failing.
• Standard Deviation of Build Times: Variability in how long builds take.
• Installation: Install ‘Build History Metrics plugin’ via Manage Jenkins → Plugins.
• Usage: Metrics appear automatically on job pages after installation and subsequent builds.

Instance-Level Metrics & Monitoring

Load Statistics

Accessible via Manage Jenkins → Load Statistics. Provides basic graphs showing:

• Build Queue Length: How many jobs are waiting for an executor.
• Busy Executors: How many executors (on controller and agents) are currently running builds.
• Response Time: Basic measure of UI responsiveness.

Useful for a quick health check and identifying chronic resource shortages.

Global Build Stats Plugin

(Formerly Hudson global-build-stats plugin) This plugin aggregates build results across all jobs on the instance and allows creating custom charts.

• Installation: Install ‘Global Build Stats’ plugin.
• Configuration: Go to Manage Jenkins → Global Build Stats. Click ‘Initialize stats’ to gather historical data. Click ‘Create new chart’.
• Chart Creation: Define chart title, size, time scale (daily, weekly), time length, and select jobs to include (or all jobs).
• Viewing: Charts display build success/failure/unstable trends over the selected period across the chosen jobs.

Monitoring Plugin (Prometheus Integration - Recommended)

For comprehensive monitoring, the ‘Monitoring’ plugin is highly recommended. It exposes detailed Jenkins internal metrics in a format compatible with monitoring systems like Prometheus.

• Installation: Install ‘Monitoring’ plugin.
• Configuration: Provides an endpoint (usually /prometheus or /metrics/<api_key>) that Prometheus can scrape.
• Metrics Exposed: Includes JVM metrics (memory, GC, threads), executor usage, agent status, queue length, job health, build durations, plugin stats, and much more.
• Visualization: Use Grafana with a Prometheus data source to create powerful dashboards visualizing Jenkins health and performance over time.
• Benefits: Provides deep insights, allows setting up alerts based on metrics (e.g., alert if queue length exceeds threshold), and integrates Jenkins into your standard infrastructure monitoring.

Choosing the Right Approach

Start with the built-in metrics and Pipeline visualizations. For more detailed job reliability insights, use the ‘Build History Metrics’ plugin. For instance-wide trends, ‘Global Build Stats’ offers simple charts. For robust, production-grade monitoring and alerting, implement the ‘Monitoring’ plugin with Prometheus and Grafana.