Azure Database for PostgreSQL Flexible Server - Database Is Alive metric stuck at 0 and Query Performance Insight stopped working after restart

Question

Azure Database for PostgreSQL Flexible Server - Database Is Alive metric stuck at 0 and Query Performance Insight stopped working after restart

Victor Danner 0

Hi everyone,

I'm experiencing an issue with an **Azure Database for PostgreSQL - Flexible Server (**PostgreSQL version: 16.14) after performing a restart (I also tried a Stop/Start, but the issue persists).

The PostgreSQL server appears to be working normally:

The application can connect without any issues.
New connections are being accepted.
Queries execute successfully.
The database is fully operational from the application's perspective.

However, I noticed that the "Database Is Alive" metric in Azure Monitor has been stuck at 0 since the restart.

At the same time, I also lost access to performance-related features in the Azure Portal:

Query Performance Insight no longer displays any data.
The "Top Queries" / most expensive queries view is also empty.

Because these features stopped working at the same time that the Database Is Alive metric became 0, I suspect they are related and that the monitoring/telemetry pipeline may not be functioning correctly.

So far, I have:

Restarted the server.
Performed a Stop/Start cycle.
Confirmed that the database is healthy and serving traffic.
Verified that the application continues to work normally.

Has anyone experienced this issue before?

Is there any known way to restart or recover the Azure monitoring components for a Flexible Server, or is this likely a platform-side issue that requires Microsoft intervention?

Any suggestions or ideas would be greatly appreciated.

Thank you!

Ganesh Chelluri 190 Reputation points Microsoft External Staff Moderator

2026-08-04T05:08:19.4133333+00:00
Hi @Victor Danner ,

Your instinct is right this looks like the telemetry side, not the engine. If the DB were really down, connections and queries wouldn't work, so "Database Is Alive = 0" points to the metric agent not reporting.

Restart and Stop/Start won't help here, there's no customer-side way to restart the monitoring components.

Two things to check first:

Are other metrics (CPU, memory, connections) also flat since the restart, or only this one?

In Server parameters, confirm pg_qs.query_capture_mode and pgms_wait_sampling.query_capture_mode aren't set to none QPI shows nothing when capture is off, and that one you can fix yourself.

If metrics are broadly missing and capture modes look fine, raise a support case with the server name, region, and restart time in UTC.
Victor Danner 0 Reputation points

2026-08-04T10:32:21.64+00:00
Hi @Ganesh Chelluri ,

Thank you for your response.

I checked the server parameters, and they are configured as follows:

pg_qs.query_capture_mode = TOP

pgms_wait_sampling.query_capture_mode = ALL

metrics.collector_database_activity = ON

I also verified that other Azure metrics are being collected and displayed correctly. CPU, Memory, Connections, Storage, and Throughput are all updating normally. The only metric that remains incorrect is Database Is Alive, which has been stuck at 0 since the restart.

Additionally, Query Performance Insight and the "Top Queries" view in the Azure Portal no longer show any data. Since these issues started at the same time, I suspect they are related to the same telemetry problem.

Unfortunately, I don't have an Azure support plan that allows me to open a technical support case, so I'm hoping someone in the community has encountered this issue before or knows of a possible workaround.
Ganesh Chelluri 190 Reputation points Microsoft External Staff Moderator

2026-08-04T17:09:14.4733333+00:00
Hi @Victor Danner ,

Check these two things:

Server parameters >

shared_preload_libraries

Confirm :

pg_qs, pgms_wait_sampling
pg_stat_statements

Are all listed. Capture modes only apply if these libraries load at startup. If one dropped off, QPI and Top Queries go blank even with capture set to TOP/ALL. Changing this needs a restart to apply, and it's fixable from your side.

azure_sys database connect and confirm it exists. Query Store data and the collector both work against it, so if it's missing, both features go quiet together.

On the metric: Database Is Alive is collector-reported, so it reflects the collector, not your engine. Your connections and queries confirm the database is healthy.

If both look fine: Raise a support request from the portal under Help + support. Include server name, region, and the UTC time of the restart. A Developer plan covers technical requests at low cost and can be cancelled after.
Ganesh Chelluri 190 Reputation points Microsoft External Staff Moderator

2026-08-06T02:45:27.0066667+00:00

Hi @Victor Danner ,

Just checking in to see if the information provided earlier was helpful. Please let us know if you need any additional details or assistance, and we’ll gladly continue supporting you.

1 answer

Your answer

Ganesh Chelluri 190 Reputation points Microsoft External Staff Moderator

2026-08-04T05:08:19.4133333+00:00

Hi @Victor Danner ,

Your instinct is right this looks like the telemetry side, not the engine. If the DB were really down, connections and queries wouldn't work, so "Database Is Alive = 0" points to the metric agent not reporting.

Restart and Stop/Start won't help here, there's no customer-side way to restart the monitoring components.

Two things to check first:

Are other metrics (CPU, memory, connections) also flat since the restart, or only this one?

In Server parameters, confirm pg_qs.query_capture_mode and pgms_wait_sampling.query_capture_mode aren't set to none QPI shows nothing when capture is off, and that one you can fix yourself.

If metrics are broadly missing and capture modes look fine, raise a support case with the server name, region, and restart time in UTC.
Victor Danner 0 Reputation points

2026-08-04T10:32:21.64+00:00

Hi @Ganesh Chelluri ,

Thank you for your response.

I checked the server parameters, and they are configured as follows:

pg_qs.query_capture_mode = TOP

pgms_wait_sampling.query_capture_mode = ALL

metrics.collector_database_activity = ON

I also verified that other Azure metrics are being collected and displayed correctly. CPU, Memory, Connections, Storage, and Throughput are all updating normally. The only metric that remains incorrect is Database Is Alive, which has been stuck at 0 since the restart.

Additionally, Query Performance Insight and the "Top Queries" view in the Azure Portal no longer show any data. Since these issues started at the same time, I suspect they are related to the same telemetry problem.

Unfortunately, I don't have an Azure support plan that allows me to open a technical support case, so I'm hoping someone in the community has encountered this issue before or knows of a possible workaround.
Ganesh Chelluri 190 Reputation points Microsoft External Staff Moderator

2026-08-04T17:09:14.4733333+00:00

Hi @Victor Danner ,

Check these two things:

Server parameters >

shared_preload_libraries

Confirm :

pg_qs, pgms_wait_sampling
pg_stat_statements

Are all listed. Capture modes only apply if these libraries load at startup. If one dropped off, QPI and Top Queries go blank even with capture set to TOP/ALL. Changing this needs a restart to apply, and it's fixable from your side.

azure_sys database connect and confirm it exists. Query Store data and the collector both work against it, so if it's missing, both features go quiet together.

On the metric: Database Is Alive is collector-reported, so it reflects the collector, not your engine. Your connections and queries confirm the database is healthy.

If both look fine: Raise a support request from the portal under Help + support. Include server name, region, and the UTC time of the restart. A Developer plan covers technical requests at low cost and can be cancelled after.
Ganesh Chelluri 190 Reputation points Microsoft External Staff Moderator

2026-08-06T02:45:27.0066667+00:00

Hi @Victor Danner ,

Just checking in to see if the information provided earlier was helpful. Please let us know if you need any additional details or assistance, and we’ll gladly continue supporting you.

Answer 1

Hi @Ganesh Chelluri ,

Thank you for your suggestions.

I verified all of the items you mentioned:

shared_preload_libraries includes:
- pg_qs
  - pgms_wait_sampling
    - pg_stat_statements
    - I already performed both a server restart and a full Stop/Start cycle.
    - The azure_sys database exists and is accessible.

During the investigation I also found the following:

query_store.query_texts continues to contain query texts.
query_store.query_plans contains 0 rows.
query_store.qs_view stopped receiving new data on 2026-05-14, immediately after the restart.
query_stats.pgms_stats_query_counts also stopped updating on the same date.
Database Is Alive has remained at 0 ever since.
CPU, Memory, Connections, Storage and Throughput metrics continue to work normally.
Query Performance Insight and Top Queries are no longer available in the Azure Portal.

These findings suggest that the PostgreSQL engine is healthy, but the internal telemetry components responsible for Query Store and monitoring stopped collecting data after the restart.

Since both Restart and Stop/Start did not restore telemetry, do you believe this points to an internal platform issue rather than a configuration issue?

Thank you for your help.