Problem Definition

When using Bitbucket self-hosted runners for Pipelines, steps can be queued for a maximum of 2.5 hours:

• https://support.atlassian.com/bitbucket-cloud/docs/runners/

When running multiple builds with multiple parallel steps that have a long duration, some steps may not be picked up by a runner because they stay as "Queued" for more than 2.5 hours while the other steps are running.

One would need to configure additional self-hosted runners so that the steps don't stay as "Queued" for more than 2.5 hours, and therefore buy more devices to use more runners.

Suggested Solution

The request is to make the step queue time in self-hosted runners configurable so that customers can increase it as needed.

Workaround

The workarounds at the moment are:

1. Setting up more self-hosted runners (which might lead to additional costs, if more machines are needed for the additional runners)

1. For Linux Docker runners specifically, the Autoscaler for Runners on Kubernetes on Kubernetes can be used to scale runners depending on the available jobs to be executed:

https://support.atlassian.com/bitbucket-cloud/docs/autoscaler-for-runners-on-kubernetes/