If you are currently on a container-based plan for CircleCI your jobs may queue if there are not enough containers in your plan to accommodate the current workload.
Job queueing can also occur on Performance, Scale, and Custom plans if there are enough builds triggered at the same time that require containers beyond the limit of containers alotted to your respective plan.
For Performance, Scale, and Custom plans please check out our Pricing page for the details of concurrency limits that apply to your account.: https://circleci.com/pricing/
Understanding container usage
As a simple example, if you have a total of 4 containers available on your account plan, you could:
- Run 4 single-container jobs at the same time
- Run a single job with 4x parallelism
- Run a Workflow with 4 parallel jobs
- Any combination of the above that does not exceed 4 containers at once.
When working with workflows, it is important to remember that the total max number of containers needed at one time will be required to be available in order to start the workflow.
For example, if in a "fan-out" workflow a single job fans-out into 4 parallel jobs, and fans back to 1 job, the entire workflow will require 4 free containers to begin.
The workflow in this example can not begin unless there are at least 4 containers available. This means if another single container job was running, even with three containers free the first job of this workflow can not begin. This is done so there are no large delays between jobs of a workflow.