azure for postgresql flexible server went down without any notice and not able to find out the reason

The Unavailable (Unplanned) event confirms that Azure detected an unexpected availability problem, but the displayed message does not identify the specific root cause; check the incident details and open a support request from Resource Health if no explanation is shown.

Determine the cause and current status

1. Open Azure portal → PostgreSQL flexible server → Resource health.
2. Open the event under Health history and record:
   • Start and recovery timestamps in UTC
   • Event/correlation ID
   • Cause and status
   • Any updates or recommended actions
3. Select Diagnose and solve problems or Troubleshooting tool from that page. Resource Health records ongoing platform and non-platform events, but a generic “unexpected problem” annotation alone cannot distinguish a hardware, networking, software, or workload failure.^1^
4. Search the portal for Service Health → Service issues → Active events. Also check the portal notification banner and Help + support → View outage details. These locations can show subscription or regional incidents associated with PostgreSQL.^2^
5. Review PostgreSQL server → Monitoring → Metrics over a window beginning at least 15–30 minutes before the outage. Correlate:
   • is_db_alive, if available
   • Active and failed connections
   • CPU and memory percentage
   • Storage percentage/free space
   • IOPS and disk queue depth
   Flexible Server metrics are emitted at one-minute intervals, although some storage-related metrics can take up to five minutes to appear.^3^

A spike in connections or memory is evidence to investigate, not proof of this outage’s cause. A similar case was resolved by reducing excessive connections and increasing memory, but that action applies only if your metrics show the same pattern.^4^

If the server remains unavailable

Allow Azure’s recovery operation to proceed unless Resource Health recommends a customer action. For unplanned failures, HA-enabled servers can fail over to their standby; without HA, Azure attempts a restart and can provision replacement compute if that restart fails.^2^

If it remains unavailable or repeatedly changes state:

1. In Resource health, select the option to contact support.
2. Alternatively, use Help + support → Create a support request and select the affected resource.
3. Include:
   • Subscription, server resource name, and region
   • Exact UTC start/end times
   • Resource Health event/correlation ID
   • Screenshots or exported event details
   • Connection errors and affected application scope
   • Relevant metric graphs
   • Whether HA is enabled
   • Whether the server recovered automatically or after an action

A support request for an unavailable resource caused by a platform event can be submitted from Resource Health without an active support agreement.^5^ Request the RCA or post-incident explanation for that specific event; PostgreSQL logs and Activity Log entries might not contain details about failures below the managed database layer.

Prevent another outage from arriving without notice

Configure both alert types because they cover different scopes:

• Resource Health alert: alerts when this specific resource changes health.
• Service Health alert: Service Health → Service issues → Create service health alert, select the subscription, PostgreSQL service, relevant regions and event types, then attach an action group for email, SMS, webhook, mobile push, or ITSM notification.^6^

Service Health alerts do not generate notifications for Resource Health events, so configuring only Service Health is insufficient.^6^

For production workloads, consider HA after recovery. Flexible Server monitors the primary and standby and initiates automatic failover when the primary is determined to be unreachable.^7^

References

1. Azure Resource Health overview - Azure Service Health | Microsoft Learn
2. Overview of business continuity in Azure Database for PostgreSQL flexible server
3. Monitor using metrics and logs in Azure Database for PostgreSQL flexible server
4. Azure Database for PostgreSQL - Flexible Server unavailable twice a day - Microsoft Q&A
5. Azure Resource Health FAQ
6. Create Service Health alerts
7. High availability in Azure Database for PostgreSQL flexible server