Why "DB Running Out of Disk Space" Alerts May Not Work on Azure PostgreSQL Flexible Server (4381204)

Users may expect comprehensive disk space alerts for cloud-managed databases, just as they get on-premises. For PostgreSQL agent, the “DB Running Out of Disk Space” rule predicts low disk conditions using tablespace growth and free space metrics. On Azure PostgreSQL Flexible Server, users notice disk-full events are not always detected by this rule.

Azure PostgreSQL Flexible Server (and AWS RDS) is database-as-a-service (DBaaS), you do not have access to the underlying virtual machine or operating system.  
This creates two key limitations:

1. Database Agent Limitations:  

• PostgreSQL functions that reveal tablespace size (like `pg_tablespace_size`) may be blocked or unavailable by the cloud provider.
• Agents cannot retrieve detailed storage stats, temp file usage, or growth rates from inside the database.

2. Infrastructure/Host Agent Blocked:

• Users cannot install or run traditional host/infrastructure monitoring agents.
• OS-level disk metrics (like “spaceAvailable”) are inaccessible.
• Database metrics are abstracted, only what Azure exposes via its API.

Therefore, the rule for "DB Running Out of Disk Space,” which needs both the growth rate of tablespaces and real host disk availability, cannot operate fully on cloud DBaaS.  
Alerts will **not trigger reliably** unless the metrics are available and exposed via native cloud monitoring.

Use Azure-native monitoring and alerting:  

• Configure Azure Monitor to track storage thresholds and receive alerts.
• Review Azure’s metrics on provisioned storage and consumption.