* Use a force resync when initialising full project history, improve error handling, increase default mongo timeout
* Don't remove history ID if error on initialisation
* Rename MAX_PROJECT_UPGRADES to MAX_UPGRADES_TO_ATTEMPT and skip upgrades in the current batch once we reach the limit
* Catch SIGINT and wait for a safe time to exit
* Set overleaf.history.upgradeFailed when errors during history initialisation
GitOrigin-RevId: 2c987926cac9c9727a7b39e9003aa8626d56423f
Extend batchedUpdate to accept find() options, and use a hint in upgrade script query to suggest mongo uses the _id index
GitOrigin-RevId: 7115f84d8be0c78ccb443051e911c931bf4aa0de