The slurmdbd process died at 1645 on Wednesday and was restarted at 0850 this morning. Slurm jobs were unaffected and continued tio run. But sacct commands would've failed overnight.
This is a retrospective status page to acknowledge the degraded service. The issue is resolved.