WAL files are not being cleaned up, leading to storage growth very rapidly in Azure database for postgres flexible server

Question

WAL files are not being cleaned up, leading to storage growth very rapidly in Azure database for postgres flexible server

Manikanteswar Punnam 40

For the last 6 days since 10th April 2025, i have been noticing the storage growth for production database (azure database for postgres flexible server) is growing very rapidly which is very unusual for our company. I am suspecting the incident which happened on this 10th April when i deleted the Read replica of our primary server, i am not sure whether that deletion happened successfully, however it is not appearing in portal. So with suspicion, i queried the WAL size and actual data size. Here are the results,

User's image

so almost half of the disk usage is WAL only currently.

This size is growing insane affecting our main storage, it will become a very big problem soon, so we need to fix this before it hit us worse. currently because of this our storage percentage reached 33% within 6 days from 34%, i.e. 32% increase in just 6 days. Current Storage capacity is 256 GB, it means it has grown close to 82 GB which is insane...

Kindly requesting you to help me solve this issue

Manikanteswar Punnam 40 Reputation points

2025-04-16T15:23:26.7133333+00:00

Correction: because of this issue, storage size reached 66% within 6 days from 34%
Saraswathi Devadula 3,155 Reputation points Microsoft External Staff Moderator

2025-04-16T15:44:26.6133333+00:00

Hello Manikanteswar Punnam

As per our discussion, could you kindly provide the business justification? This will assist us in conducting a more thorough investigation into the issue.
Manikanteswar Punnam 40 Reputation points

2025-04-16T15:55:05.54+00:00

I am urging you to solve this issue quickly because this issue has direct and critical business impact in the following ways:

Escalating Infrastructure Costs: The excessive and unexpected storage consumption has resulted in increased Azure costs. Since storage in Azure Flexible Server auto-scales with usage, we are being charged for space that should ideally be freed up automatically.

Risk of Service Disruption: If WAL files continue to accumulate unchecked, we risk hitting storage thresholds that may lead to performance degradation or even a crash of our production database, which powers key customer-facing features. This poses a serious risk to business continuity.

Operational Overhead: Our engineering team is currently forced to manually monitor and manage storage usage and consider manual scaling. This diverts valuable engineering time away from delivering business features and increases operational burden.

Scalability Concerns: Without a resolution, we may be forced to scale the server instance prematurely, incurring further costs and complexity — despite the actual application data not growing at a matching rate.
Saraswathi Devadula 3,155 Reputation points Microsoft External Staff Moderator

2025-04-16T15:58:57.2666667+00:00

I appreciate you providing these details. I'm working internally and will do my utmost to address and resolve this issue.
Saraswathi Devadula 3,155 Reputation points Microsoft External Staff Moderator

2025-04-16T17:34:47.8433333+00:00

Hello Manikanteswar Punnam

Please be informed that we worked internally and tried to restart the server, and we request you to retry the read replica and check the storage.
Manikanteswar Punnam 40 Reputation points

2025-04-16T17:36:06.0166667+00:00

I just noticed that storage came down to actual data size. Issue is resolved, now. Thanks a lot Sarawathi for the immediate action.

Accepted answer

0 additional answers

Your answer

Manikanteswar Punnam 40 Reputation points

2025-04-16T15:23:26.7133333+00:00

Correction: because of this issue, storage size reached 66% within 6 days from 34%
Saraswathi Devadula 3,155 Reputation points Microsoft External Staff Moderator

2025-04-16T15:44:26.6133333+00:00

Hello Manikanteswar Punnam

As per our discussion, could you kindly provide the business justification? This will assist us in conducting a more thorough investigation into the issue.
Manikanteswar Punnam 40 Reputation points

2025-04-16T15:55:05.54+00:00

I am urging you to solve this issue quickly because this issue has direct and critical business impact in the following ways:

Escalating Infrastructure Costs: The excessive and unexpected storage consumption has resulted in increased Azure costs. Since storage in Azure Flexible Server auto-scales with usage, we are being charged for space that should ideally be freed up automatically.

Risk of Service Disruption: If WAL files continue to accumulate unchecked, we risk hitting storage thresholds that may lead to performance degradation or even a crash of our production database, which powers key customer-facing features. This poses a serious risk to business continuity.

Operational Overhead: Our engineering team is currently forced to manually monitor and manage storage usage and consider manual scaling. This diverts valuable engineering time away from delivering business features and increases operational burden.

Scalability Concerns: Without a resolution, we may be forced to scale the server instance prematurely, incurring further costs and complexity — despite the actual application data not growing at a matching rate.
Saraswathi Devadula 3,155 Reputation points Microsoft External Staff Moderator

2025-04-16T15:58:57.2666667+00:00

I appreciate you providing these details. I'm working internally and will do my utmost to address and resolve this issue.
Saraswathi Devadula 3,155 Reputation points Microsoft External Staff Moderator

2025-04-16T17:34:47.8433333+00:00

Hello Manikanteswar Punnam

Please be informed that we worked internally and tried to restart the server, and we request you to retry the read replica and check the storage.
Manikanteswar Punnam 40 Reputation points

2025-04-16T17:36:06.0166667+00:00

I just noticed that storage came down to actual data size. Issue is resolved, now. Thanks a lot Sarawathi for the immediate action.

Answer 1

Hello Manikanteswar Punnam

Issue : Initially, you were unable to create the read replica and suspecting that due to this only the storage is increased.

Cause: VM agent is in unhealthy state due to which you are unable to create the read replica, and the storage was increased.

Solution: In further investigation, we tried to restart the server, and we request you to retry the read replica and check the storage.

later your confirmed that you are able to create the read replica and the issue is resolved.

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. Thank you for helping to improve Microsoft Q&A!

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

User's image

Share via

WAL files are not being cleaned up, leading to storage growth very rapidly in Azure database for postgres flexible server

0 additional answers

Your answer