migration 0129 takes way too long to run now on production data. Mitigation steps must be taken to allow the v2->v3 migration to complete in a timely fashion.
I recommend:
move validation_state_member and validation_result tables aside to a separate database, and truncate those tables in the db backup being restored to the main database
after the api server is back up and running, perform the schema migrations offline, then load the new data into the production database.
The unfortunate side effect here is that validation results will be unavailable for all existing devices until the process is completed. However, new device reports and validations will continue to be processed, and a device's existing status can still be seen because validation results are copied into device.health.
migration 0129 takes way too long to run now on production data. Mitigation steps must be taken to allow the v2->v3 migration to complete in a timely fashion.
I recommend:
The unfortunate side effect here is that validation results will be unavailable for all existing devices until the process is completed. However, new device reports and validations will continue to be processed, and a device's existing status can still be seen because validation results are copied into device.health.