Closed brick-green closed 1 year ago
Majority of errors are of type 400: Bad Request. The GAEN documenation (https://github.com/google/exposure-notifications-verification-server/blob/main/docs/api.md#apiissue) mentions several different reasons for why a 400 status might be returned as evident in this table e
Errors do not appear to be related to message content as there are no obvious deficiencies in failed message. Need to investigate full JSON response to see if this is helpful.
When I have a chance, I will take a look at this ticket with James.
Need to add logic to GAENTransport.kt to include error returned in JSON in logs. ENCV server docs provide context and guidance for this. Relevant portion of code is between lines 238-258 of GAENTransport.kt
In ticket 7615. I added a response errorCode of 400 Bad request Status code as @GilmoreA6 mentioned above. The errorCode will tell us a specific failure. Therefore, we can address it.
Every wa-gaen test that failed in last mile failure. They all failed with 400 Bad requests of invalid_date = 'The provided test or symptom date, was older or newer than the realm allows. For the date to be valid the illness_onset_date (or symptomDate) and date_result_release (or testDate) must be within 29 days from the day of submission. And, the illness_onset_data <= data_result_release data. For example, date of submission is 2022-12-15 the date_result_release must be >= 2022-11-17 AND <= 2011-12-15 AND the illness_onset_date >= 2022-11-17 AND <= date_result_release. Otherwise, the date considers being older.
Need to find out how RS generate the GAEN message for WA-GAEN.
Problem statement
We currently have a receiver for Google/Apple Exposure Notification (GAEN). The transport regularly fails. We need to understand why it fails and plan for a solution.
What you need to know
Failures can be reviewed on last mile problems page: https://reportstream.cdc.gov/admin/lastmile
Acceptance criteria
To do