Beginning around 12:00 UTC on the 13th of June, Engineers were alerted of elevated queue levels for Connector services in CS-US1.
A leader node was found to have higher than usual disk activity which prevented optimal job execution for Connector services.
Engineers scaled the number of Connector instances to correct the issue temporarily. Additionally, Engineers skipped long-running and duplicate jobs to help recover.
System configuration adjustments have been made to prevent these issues moving forward.