Record Locks during Patch request linked to indeterminate/silent fails from Sync Robot

[CritiqalPete]

The Sync-Robot requests include an update/patch to the Attendance object. See attached. It appears that some of these fail due to a record lock issue. Observations:

• When a sync is run a second time, some updates are completed
• Reducing the number of requests per batch and/or adding a delay between requests is not yet clearly correcting the issue
• The Robot is not picking up a failed/incomplete response from the API which would be valuable to tracking this condition Expectations:
• The API should support the specified request payload and frequency
• The API should include a response that will help flag and manage these issues in the future

Needs:

• Why is this record lock occurring? (presumably a Salesforce behavior)
• Is the API performing to spec?
• What, if any, changes need to be made on the client/request side to avoid this?
• What messaging/response needs to be identified by the client (Sync-Robot) to detect and respond accordingly? Does something need to be added to the API to support this? API_record_lock_logentry.txt

[matthewjerome]

high-level observations when running many requests to the api/coach/[coachId]/teamseasons/[teamSeasonId]/sessions/[sessionId]/attendances endpoint

• when running 20 attendance records per request with a timeout of ~200ms between requests, the error Pete pasted in the log begins to appear after the first handful of requests are run
• when running 100 attendance records per payload with a timeout of ~5000ms between requests, the error doesn't appear to occur