- Lead time: The amount of elapsed time it takes for a code commit to successfully deploy to production.
- Deployment frequency: A measure of how often a team successfully releases new updates to their users.
- Change failure rate: The percentage of deployments that result in a failure or require immediate rollback in production.
- Mean time to recovery: The average time it takes to restore a service after an unexpected failure or outage occurs.
Teams configure Azure DevOps analytics queries and Azure Monitor dashboards to visualize these indicators. Analyzing these results helps teams identify bottlenecks and prioritize where to make process improvements.
Configuring Azure Monitor and Analytics
Tracking these metrics requires setting up proper data collection and log analysis tools. Using Log Analytics workspace insights, teams can run queries to find trends in their deployment data and identify slow workflows. Enabling log query auditing helps teams find and rewrite inefficient queries that might otherwise slow down their monitoring dashboards.
For applications running in Kubernetes, monitoring must cover every layer of the environment. Teams use Container insights alongside Azure Monitor managed service for Prometheus to gather metrics from the network, cluster, and application layers. This data flows into centralized dashboards, allowing teams to build proactive alerts and identify system anomalies before they impact users.
Managing delivery metrics also requires balancing performance, cost, and security across all monitoring tools. High-volume logging can quickly become expensive, so teams should configure cost optimization settings to filter out unnecessary telemetry data. This ensures the team only stores the high-value logs needed to calculate their core delivery metrics.
For the data tier, teams must monitor performance and security without introducing overhead. A tool like Database watcher helps track database health, while recommendations from SQL Database Advisor guide teams on how to secure and optimize their databases. This end-to-end monitoring ensures that both the delivery pipeline and the underlying database systems remain fast, secure, and cost-effective.
StudyCerts
StudyCerts