Experiencing data loss for network and dependency data for VM Insights and ServiceMap
The customers in East Japan and East US have experienced data latency and data loss for the network and dependency data. VMComputer, VMProcess, VMBoundPort, and VMConnection tables in the Log Analytics workspaces were affected. This would have impacted the Map view of VM insights and the alerts based on the data collected in these tables.
Incident duration:
East Japan: Between 13 Jul 2021 00:00 UTC and 20 Jul 2021 00:24 UTC
East US: Between 17 Jul 2021 00:01 UTC and 19 Jul 2021 21:21 UTC
Root Cause: We have identified a configuration mismatch in a recent deployment affecting Service Map's ability to communicate with its underlying storage accounts. The deployment, in part, updated our method of authenticating to the storage accounts on the back-end, but a subset of accounts was not included in the update.
Mitigation: We have rolled back the deployment to resolve the configuration mismatch and recovered the alerts/logs data with minimal loss. The issue is mitigated.
We understand that customers rely on Azure Monitor VM Insights as a critical service and apologize for any impact this incident caused.
