Resolution: Won't Do
Fix Version/s: None
In the previous deployment (
GPII-3968), the performance of the preferences endpoint was not good enough to pass the existing smoke tests. The work to address these performance problems has now been merged and tested in my dev cloud.
The QA deployment can happen whenever ops is ready. There are no database changes or other migration steps required in order for QA to take place.
The production deployment is contingent on the QA review. Once that is complete, Tony is the developer responsible, and is generally available (all times UTC+2):
W: 9:30-19:00 (in meetings from 16:00, but available to chat)
We should agree on a time for the deployment at least a day ahead of time.
We need to confirm whether the telegu dataset has already been purged from the production persistence layer. If not, how will that be done and by whom?
- Find the node of a preferences pod using the web interface.
- Attach to the node using a command like:
- Attempt a valid prefs GET using a command like:
- You should see a prefs payload.
- Attempt an invalid prefs POST using a command like:
- You should see a validation error like:
Submit a reversion pull against gpii-infra that references the previous version of the universal docker container.
This work does not change the UI at all. We are mainly considered with side effects, i.e. verifying that reading and writing settings works as expected.