@@ -525,6 +525,26 @@ There is an ongoing [discussion](https://gitlab.com/gitlab-org/gitlab/-/issues/4
We will try to avoid this additional table, if not this will be partitioned based on the created timestamp with a retention period similar to BBM framework.
##### Decision: Not storing logs in the database
In BBM, `batched_background_migration_job_transition_logs is used to store the state transition logs, but we decided not to store them in the database for background operations monitoring.
Instead, we will use other types of metrics:
-**Prometheus** for time-series metrics
-**Kibana** for log aggregation and analysis
This approach reduces maintenance overhead associated with managing large log tables in the database, including:
- Partitioning and retention policies
- Index optimization
- Data cleanup procedures
- Database growth and storage requirements
These tools are purpose-built for handling high-volume metrics and logs, providing better scalability and operational simplicity compared to database tables.
See [gitlab-org/gitlab!218794](https://gitlab.com/gitlab-org/gitlab/-/merge_requests/218794) for the decision discussion.
### Considerations
- Preference should be given to organization specific operations than the cell local ones, to reduce the waiting time on user facing operations.
