Skip to content

Remove job token on completion

What does this MR do and why?

Related to #349877

It removes the job token after job completion because the token is not used after a job finishes and it's getting too expensive to maintain. In one year, the size of the index for this column has doubled in size:

# Now

List of relations
 Schema |                Name                | Type  |  Owner  |   Table   |  Size  | Description 
--------+------------------------------------+-------+---------+-----------+--------+-------------
 public | index_ci_builds_on_token_encrypted | index | joe_309 | ci_builds | 250 GB | 
(1 row)

# 1 year ago:

gitlabhq_production=> \di+ index_ci_builds_on_token_encrypted
                                        List of relations
 Schema |                Name                | Type  | Owner  |   Table   |  Size  | Description
--------+------------------------------------+-------+--------+-----------+--------+-------------
 public | index_ci_builds_on_token_encrypted | index | gitlab | ci_builds | 128 GB |
(1 row)

https://console.postgres.ai/gitlab/gitlab-production-ci/sessions/14196/commands/49793

And for partitioning the ci_builds table we need to add the partition_id column in the index's definition, but it proves too difficult to do due to the amount of data that we need to index.

Screenshots or screen recordings

Screenshots are required for UI changes, and strongly recommended for all other merge requests.

How to set up and validate locally

Numbered steps to set up and validate the change are strongly suggested.

MR acceptance checklist

This checklist encourages us to confirm any changes have been analyzed to reduce risks in quality, performance, reliability, security, and maintainability.

Edited by Marius Bobin

Merge request reports

Loading