Optimising pipeline concurrency — quick notes
Concurrency — quick notes
| Fact | Detail |
|---|---|
| Service unit | Parallel job = organization capacity for one actively running agent job |
| Consumed by | Running matrix legs and running jobs from other organization pipelines; not queued legs, steps, stages, waits, or artifacts |
| Agent versus entitlement | Agents supply execution infrastructure; parallel-job entitlement limits simultaneous service use |
| Throttle a matrix | maxParallel |
maxParallel validity | Only with matrix |
parallel + matrix | Mutually exclusive |
| Self-hosted scaling | VMSS agents are customer-managed; Managed DevOps Pools is fully managed; both still need self-hosted parallel-job entitlement |
Trade-off: more capacity can reduce queueing or parallelizable wall-clock time; cost depends on the agent and billing model.
Sources: https://learn.microsoft.com/en-us/credentials/certifications/resources/study-guides/az-400 · https://learn.microsoft.com/en-us/azure/devops/pipelines/licensing/concurrent-jobs · https://learn.microsoft.com/en-us/azure/devops/pipelines/yaml-schema/jobs-job-strategy · https://learn.microsoft.com/en-us/azure/devops/pipelines/process/phases · https://learn.microsoft.com/en-us/azure/devops/pipelines/agents/github-hosted · https://learn.microsoft.com/en-us/azure/devops/pipelines/agents/agents · https://learn.microsoft.com/en-us/azure/devops/managed-devops-pools/overview · https://learn.microsoft.com/en-us/azure/devops/managed-devops-pools/migrate-from-scale-set-agents