[tfeltin]

Description

• Lower deployment footprint: fewer containers, workers run as threads in one process, docker compose deployment goes from 8 worker services to 5. Helm unchanged
• Smaller images: removed unecessary torch, numba/umap, litellm imports, clickhouse-sqlalchemy and the full MCE package from the images that don't need them. The memory requirement is now 6 GB.
• Fixes:
  • The mce worker no longer processes one session at a time.
  • The thread-pool cap is gone.
  • The ClickHouse pool warnings are gone.
  • Neo4j deadlocks are retried.
  • The intelligence worker no longer fails on the concerns validation error.
• Insight templates for cognitive failures, one per session and one per group.
• Docs and CHANGELOG updated.

Type of Change

• Bugfix
• New Feature
• Breaking Change
• Refactor
• Documentation
• Other (please describe)

Checklist

• I have read the contributing guidelines
• Existing issues have been referenced (where applicable)
• I have verified this change is not present in other open pull requests
• Functionality is documented
• All code style checks pass
• New code contribution is covered by automated tests
• All new and existing tests pass

[Castorche]

LGTM