bitbucket shared / output variable limit

Ben Bewick

Regarding the limit of 50 shared / output variables per pipeline, I assumed this would be 50 unique named variables within the pipeline.

I currently have about 26 uniquely named variables, with many of these are reused / over written between different deploy environments.

This was working when fine when I had 2 deploy environments setup for testing.

When I added a third environment for production, it is now complaining that I have more than 50 variables used.

I had assumed that by over writting I would stay within the 50 limit, but apparently not. Is this the expected behaviour ??

Viswanathan Ramachandran

Hi @Ben Bewick

Welcome to the community.

Yes, this is the expected behavior. The 50-variable limit on Bitbucket Pipelines output-variables applies cumulatively across all steps in the entire pipeline execution, not per unique variable name. Overwriting a variable in a subsequent step still increments the counter.

To bypass this limit:

Use an artifact file (.env or .sh): Write variables to a file in one step, export it under artifacts:, and source it in subsequent steps.

Use Deployment Variables: Move environment-specific static configs into Repository Settings > Deployments under each respective environment.

Bundle values: Combine related variables into a single JSON object or string.

Hope it helps and clarifies.