Migration plan

Whenever we're not in freeze

1. Install Timescaledb on db-datanommer01
2. Extract the datanommer source code somewhere on db-datanommer01
3. Run the migration script
4. While the script is running, deploy the new apps in OpenShift
5. Unsubscribe the datanommer AMQP queue from everything and purge it
6. Stop the new datanommer pod (with oc scale --replicas=0)
7. Run the migration script again, multiple times if necessary, until it completes in a few minutes.

Start of the downtime window

1. Re-subscribe the datanommer AMQP queue to everything
2. Stop the current datanommer service
3. Run the migration script again
4. Start the new datanommer pod
5. Update the proxies to point to the new datagrepper app in Openshift
6. Merge the Badges PR
7. Run the ansible playbook to update Badges to use the backported datanommer.models file

End of the downtime window