Azure Database for PostgreSQL Flexible Server — CPU, Memory, and Storage metrics still missing after diagnostic reset

Question

Hello,

We are still experiencing the same issue with our Azure Database for PostgreSQL Flexible Server where CPU, memory, and storage metrics are not being collected. All other metrics are still flowing correctly.

It looks like this:

What we already tried (based on prior advice):

• Reset diagnostic settings: turned off “send to Log Analytics Workspace”, saved, waited, and re-enabled with all metric categories selected → no effect.
• Verified Log Analytics: confirmed retention and data caps are fine; other metrics are being ingested into the AzureMetrics table as expected.
• Checked queries: using correct schema/table (AzureMetrics), and only CPU, memory, and storage metrics are missing.
• Resource Health: no ongoing issues reported for our PostgreSQL Flexible Server resource.

Additional context:

• Diagnostic settings are configured to send all metrics to Log Analytics.

Relevant PostgreSQL server configs:

pg_qs.query_capture_mode = ALL   metrics.pgbouncer_diagnostics = ON   metrics.collector_database_activity = ON   metrics.autovacuum_diagnostic = ON  

• This exact issue happened a few months ago, but metrics reappeared automatically after a few days. Now, it has been several weeks, and the metrics remain unavailable.

Questions:

1. Is there a known issue with CPU, memory, and storage metric collection for Flexible Server?
2. What further actions can we take to restore full metric visibility (beyond diagnostic reset)?

Any guidance would be appreciated. Thank you!

Answer 1

Hi Serhii Boiko,

Based on our private conversation, we’ve confirmed that Azure Monitoring metrics are currently not showing up for Azure Database for PostgreSQL Flexible Server in the Azure portal.

Upon investigation, we found that the component that collects and forwards server metrics (Fluentd was identified as unhealthy ) (CPU, Memory, Storage) stopped working properly, which caused those metrics to go missing. Since the component responsible for collecting logs was not functioning correctly, the logs were not being generated. The problem has now been mitigated by restarting the container from backend.

I hope this has been helpful!

Your feedback is important so please take a moment to accept answers. If you still have questions, please let us know what is needed in the comments so the question can be answered.

If this answers your query, please click Accept Answer and Upvote for was this answer helpful.

Thanks!

Kalyani