Systems Issues Knowledge Base

SQL Recoverable Error Closed Connection

starting in April 2015, we began experiencing intermittent errors in PRD as reported by Splunk.  Click here to see sample email notification

These errors were captured in Splunk but had no detectable impact on users.  We were unable to identify a systemic or application change that occurred at the same time so the root cause was a bit of a mystery. The additional factor complicating discovery of the root cause was the lack of discernible pattern in the occurrences.  The one thing that seemed to be consistent was that the majority of them seemed to occur at times of heavy database use, specifically during batch.  The DBA's were unable to identify any issues on the database side. Given the fact that the issues were not user facing and we were in the midst of several competing initiatives, we were unable to devote focused effort to resolving.

In early May 2015, Brett made the connection between (2) KFS Application parameter that he'd changed in an effort to resolve some user reported issue.  The parameter changes were performed the very day in early April that we saw the first occurrences of the error.  When Brett made the connection between the errors and the parameter changes, he returned the parameters to their original values. Since that time, the issue has not recurred.