Publishing Metrics
The Metrics page provides insights into the performance and resource
usage of your project's deliverables. It provides detailed data for your project's deliverable
activities over the last 30 days. The page is organized into several sections, including a
summary of the deliverables and detailed metrics for runs, queue times, durations, and disk
requirements.
Note: To adjust the number of days that metrics are
recorded, an administrator can set the
executor.service.metrics.disk.store.time
property for the Task Executor
container. This property determines the length of the time window for which metrics data is
stored and displayed.Deliverables Summary
- Total Runs - The total number of times the deliverable builds have been executed for a project within the last 30 days.
- Queued P50 - The 50th percentile of time that builds spend in a queue, meaning 50% of the builds waited in the queue for this amount of time or less.
- Duration P50 - The 50th percentile of the total duration of builds, meaning 50% of the builds completed within this time or less.
- Used Disk Space - The total disk space consumed by artifacts (build outputs) and repositories.
Deliverables Metrics
Runs:
- Successful - The number of deliverable builds that completed without errors.
- Failed - The number of deliverable builds that did not complete successfully due to errors.
- Memory issues - The number of deliverable builds that failed
due to 'Out of Memory' errors.Note: If you encounter memory issues, contact your platform administrator to increase the memory allocated for the Task Executor container and OPE. This ensures there are sufficient resources for processing tasks. The OPE memory can be adjusted using the
executor.service.ope.max.memory
property. - Success Rate - The percentage of deliverable builds that were successful out of the total attempted runs.
Queued:
- Min - The shortest time spent for a deliverable build in the queue.
- Max - The longest time spent for a deliverable build in the queue.
- Average - The average time spent for a deliverable build in the queue, calculated across all runs.
- P90 - The 90th percentile of time that deliverable builds spent
in the queue, meaning 90% of the builds waited in the queue for this amount of time or
less.Note: If the Average or P90 metrics are too high, the platform administrator may reduce the queue time by increasing the number of parallel builds using the
executor.service.parallel.publishing.builds
property.
Duration:
- Min - The shortest time taken to build a deliverable.
- Max - The longest time taken to build a deliverable.
- Average - The average time taken to build a deliverable, calculated across all runs.
- P90 - The 90th percentile of time taken to build a deliverable, meaning 90% of the runs completed in this time or less.
Disk Requirements:
- Repository Clones Size - The total disk space used by the repositories.
- Artifacts Size - The total size of all artifacts (build outputs) generated by the deliverables.
Utilizing Metrics Data
The metrics provided can help you:
- Monitor performance trends and identify potential bottlenecks.
- Assess the reliability of your deliverable runs via success rates and failure counts.
- Make data-driven decisions to streamline your publishing process.