• Type: Deployment
    • Status: Resolved
    • Priority: Major
    • Resolution: Done
    • Fix Version/s: None
    • Component/s: universal
    • Labels:



      This is a new ticket for the work originally discussed in GPII-4006 and earlier tickets, which covers adding validation to flow manager and preferences endpoints, and also reducing the memory required for validation under heavy load. In addition, the work to remove the context manager described in GPII-3117 is included. This latest pull also includes the fixes for problems revealed in previous QA efforts, namely GPII-4256 and GPII-4258.

      Information for OPs

      When is a good time for you and the Ops team to deploy this change?

      The QA deployment can happen whenever ops is ready, ideally on Monday, December 16th. No database changes or other migration steps are required in order for QA to take place.

      How can we verify that the change is working correctly in production?

      Command Line Test of Validation Infrastructure

      1. Find the node of a preferences pod using the web interface.
      2. Open a rake shell, i.e.
        rake sh
      3. Attach to the node using a command like:
        kubectl --namespace gpii exec preferences-899b5bbf-xswr7 -it -- sh
      4. Attempt a valid prefs GET using a command like:
        curl http://localhost:8081/preferences/jaws
      5. You should see a prefs payload.
      6. Attempt an invalid prefs POST using a command like:
        curl -X POST -H "Content-Type: application/json" --data '{ "foo": "bar" }' http://localhost:8081/preferences
      7. You should see a validation error like:
        "isError":true,"message":"Your request was invalid.  See the errors for details.","isValid":false, // trimmed ...

      QA testing using Morphic Client

      The Morphic client will be tested against staging to confirm that an auto-keyed-in user can change and save their preferences to the cloud. In previous testing, there were problems with language and contrast settings, at a minimum I will confirm that these are working. We will go through a more extensive QA effort with the same container version in my dev cloud.

      What is the rollback plan if things don't go well?

      Submit a reversion pull against gpii-infra that references the previous version of the universal docker container.

      Information for QA

      I will deploy the updated container version to my dev cloud, we should perform a preliminary pass at QA there. Kavya to provide documentation regarding standard QA steps.




            the-t-in-rtf Tony Atkins
            the-t-in-rtf Tony Atkins
            0 Vote for this issue
            1 Start watching this issue