Increase max_total_records per new Spanner mutation limit

[data-sync-user]

Spanner’s doubled its mutation limit from 20k to 40k:

https://cloud.google.com/spanner/docs/release-notes#September_27_2022

https://cloud.google.com/blog/products/databases/cloud-spanner-doubles-the-number-of-updates-per-transaction

This allows us to increase the max_total_records setting for batch uploads. We should take advantage of this as it will reduce the number of requests needed by clients to make larger writes. This involves:

• Likely doubling the existing limit – we can double check how the max total mutation count is calculated with the notes in BATCH_COMMIT.txt
• Increase the limit via the spanner_config.ini file
  • We should double check if we can kill this file now, it’s only needed to workaround a settings issue. If it’s no longer necessary we could easily increase this limit via SRE changing an env var instead
• Roll the change out to solely canary and monitor Spanner’s metrics for any adverse changes
  • Log an issue with firefox-ios to update their hardcoded value for this setting (and we might as well give them a PR, it’s a 1 line change) The pending (as of 2023-05-30) SyncManager work will remove their hardcoding of values: they'll begin reading them from the server

┆Issue is synchronized with this Jira Task

[pjenvey]

Reopened as data-sync-used seems to have closed this by mistake