Resource class pinning for low-traffic workload cost reduction

Compute Cost Optimization

For low-traffic projects, utilizing the smallest available resource_class can significantly reduce monthly compute spend compared to the default executor settings. To ensure consistent billing, the resource_class must be explicitly pinned in the .circleci/config.yml file rather than relying on project-level defaults.

While pinning to a smaller runner (such as medium) reduces costs, there is uncertainty regarding the interaction between explicit resource pinning and the auto_cancel feature when managing concurrent runs for trivial tasks.

• How does explicit resource class pinning affect the cost calculation when auto_cancel triggers a job termination?
• What is the recommended verification method to ensure a job is not over-provisioned when using an orb that defines its own default resource class?