Investigate metropolitan reingestion workflow failures/time outs

[stacimc]

Description

The Metropolitan reingestion DAG take an extremely long time. It is a weekly DAG and historically times out at the end of the week having only completed 6 or 7 reingestion days, and then starts over. This means reingestion is not getting completed on the proper schedule, and the DAG is perpetually running (so it is almost always stopped and forced to error when a deploy runs).

We need to investigate why the DAG takes such a long time and consider:

• are there more efficient ways to reingest?
• can we alter the reingestion schedule?
• can we make the DAG resilient to being shut down during a deploy (similar to how batched updates work)?

This issue is for investigating the problem. We may need to open new issues for work, depending on our findings.

Additional context