Update Readd Pages Background Job

[ikreymer]

The re-add pages background job should be updated in the following ways, to ensure all pages are in the right format for use with #2347.

Based on issues running migration with background jobs, here's a list of things that should be improved:

• Readd pages job should have more memory to avoid OOM errors - perhaps as much as backend pod
• Should handle existing data from QA review, perhaps either upserting, or storing the data separately and readding after import?
• Handle any WACZ loading errors and retry
• Handle duplicate pages somehow: if the same WACZ is uploaded multiple times, generate a unique id for the page and store page id in a pageId field?
• Other issues observed: distinct() in mongo fails due to large size, not sure if there's a good solution yet.

[ikreymer]

Jotting down some more impl thoughts:

• create a temp collection for any pages for crawl / org qa_data_<crawl_id> for any pages that have approved,notes,userid,modified, then readd after pages reimported. could be done with two aggregations, i think.
• Uploads avoid duplicate pages by always using new page id, via Upload Fixes: #2397

[tw4l]

Related PR #2400 moves re-adding pages from background jobs into the migration to avoid later migrations running before pages have been re-added.

[ikreymer]

Addressed via #2400, #2406