Maximum number of concurrent agents that can be provisioned from this template.

Set to 0 for unlimited (still subject to the cloud-level limit). Use this to cap how many agents a single template can consume, for example to stay within GitHub Actions concurrency limits for a specific runner type.