Clean up some nightly build infrastructure cruft

[zaneselvans]

Overview

Closes #3927
Closes #3643
More generous timeouts/retries may address #3958

Tasks

Preview Give feedback

1. Bump micromamba Docker image to v2.0.3
2. Remove obsolete GCE_INSTANCE and pudl-deployment-tag/pudl-deployment-dev refs
3. Update docs to reflect use of Google Batch and rebuild the docs to test.
4. Update link to RTD backend in our stable release checklist.
5. Run full branch deployment with obsolete references to GCE_INSTANCE etc removed.
6. Run fast branch deployment with simplified logic (as suggested by @jdangerx in Don't deploy outputs when using workflow_dispatch trigger on main #3643)
7. Remove parquet distribution steps
8. Test GCS & S3 distribution in workflow_dispatch, and then remove the distributed files.
9. Run fast branch deployment with test distribution of outputs to GCS & S3 enabled for workflow_dispatch
10. Use gcloud storage instead of aws s3 for S3 distribution
11. Remove AWS CLI from the docker container
12. Run FAST test deployment using gcloud storage instead of aws s3
13. Remove S3 and GCS test outputs at /$BUILD_ID
14. Switch back to automatically removing test outputs as soon as they've been deployed on GCS / S3
15. Switch back to doing FULL rather than FAST etl
16. Turn integration & validation tests back on
17. Make VCE RARE row count asset check blocking again
18. Do a self-review of the PR and call out any questions or issues.
19. Do a final FULL branch deployment to make sure everything is actually still working
20. Update to using Postgres v16 and the Ubuntu 24.04 micromamba image

[bendnorman]

Looks good! Just a few small suggestions.