refactor where we send mlbf success ping to statsd

[eviljeff]

Right now, because we fire a separate celery task to carry out the mlbf/stash upload to remote settings, we send the success ping to statsd after we successfully build a bloom filter, but because the upload happens. Meaning if there is a problem with the upload there is no indication in grafana - apart from the lack of the upload pings, but that can correctly occur when no upload is needed.

Options:

• move the success ping inside upload_filter
• send two additional pings, when we start upload and when we end, and add a grafana alert when both don't occur
• wait for the celery task to complete before we continue execution of the cron job/management command.

Note, the celery task was a deliberate change (https://github.com/mozilla/addons-server/pull/13909) because only celery workers have the required access to upload files to remote settings.

┆Issue is synchronized with this Jira Task

[KevinMind]

Old Jira Ticket: https://mozilla-hub.atlassian.net/browse/ADDSRV-500