Investigate lazy loading of ResultSet in PatientPrimaryPhoneDataLoader

[emmastephenson]

Background

One of our prod containers crashed on 7/7/22 because there were a number of DataFetcherExceptions returned from a database query. The root cause of this was that we were trying to send too many bind variables with a single query, and specifically trying to request ResultSet when using the PatientPrimaryPhoneDataLoader.

Action requested

• Research why ResultSet is being requested as part of the PatientPrimaryPhoneDataLoader at all
• See if this is related to a single graphQL query (my guess is some kind of test result list query, because there were many /testResults[1234]/patient/telephone queries all back to back.)
• Either break up the initial query to request less information (may not be feasible) or stop lazy-loading the ResultSet for patient phone number

Acceptance Criteria

[guessing here] We can successfully execute a query to get the test result list for an organization that has over 15,000 testEvents without hitting a postgreSQL exception.

Additional context

StackOverflow about the cap on data binds: https://stackoverflow.com/questions/49274390/postgresql-and-hibernate-java-io-ioexception-tried-to-send-an-out-of-range-inte
pgjdbc Github post about the same issue: pgjdbc/pgjdbc#1311
Our container log stack trace: https://gist.github.com/emmastephenson/cba394abeed60980ae1313c5bddd46a2
Original page about long-running database queries: https://prime-cdc.pagerduty.com/incidents/Q1ZAL63CY41CDZ