Azure Managed Redis - REST API to flush

Question

Azure Managed Redis - REST API to flush

Eddy Lynch 60 Microsoft Employee

Hello,

I am using Azure Managed Redis (redis enterprise), and I tried using the Azure management rest api "Microsoft.Cache/redisEnterprise/<cluster-name>/databases/default/flush"

I get a successful result on the operation. However, I notice the cache is still at 2GB used memory

Eddy Lynch 60 Reputation points Microsoft Employee

2026-04-28T13:57:38.7333333+00:00

Hi @Pilladi Padma Sai Manisha thanks for the response. Just to confirm, that this also applies for Azure Managed Redis? Because, when I did the flush, I also noted that the stored keys metric did not drop also after quite some time
Pilladi Padma Sai Manisha 8,410 Reputation points Microsoft External Staff Moderator

2026-04-28T14:11:58.3033333+00:00

Yes,this applies to Azure Managed Redis as well.

If the stored keys metric hasn’t dropped, the first thing to check is whether the flush operation has actually completed. The API is asynchronous, so it’s important to confirm the Azure-AsyncOperation (or Operation-Location) shows Succeeded.

That said, even after a successful flush, Azure metrics (including key count) can take a couple of minutes to update. So a delay there is normal.

The quickest way to confirm is to check at the Redis level,if DBSIZE returns 0, then the flush has worked and what you’re seeing is just metric lag. If it’s not 0, then the operation likely hasn’t completed or didn’t apply to the expected scope.

If you’re using geo-replication, you’ll also want to make sure the flush was done at the group level.

Happy to take a closer look if you can share what DBSIZE returns.
Pilladi Padma Sai Manisha 8,410 Reputation points Microsoft External Staff Moderator

2026-04-30T15:25:20.94+00:00

Hi
I hope you had a chance to review the information shared earlier, and I hope this information has been helpful! If you still have questions, please let us know what is needed in the comments so the question can be answered.

Answer accepted by question author

0 additional answers

Your answer

Eddy Lynch 60 Reputation points Microsoft Employee

2026-04-28T13:57:38.7333333+00:00

Hi @Pilladi Padma Sai Manisha thanks for the response. Just to confirm, that this also applies for Azure Managed Redis? Because, when I did the flush, I also noted that the stored keys metric did not drop also after quite some time
Pilladi Padma Sai Manisha 8,410 Reputation points Microsoft External Staff Moderator

2026-04-28T14:11:58.3033333+00:00

Yes,this applies to Azure Managed Redis as well.

If the stored keys metric hasn’t dropped, the first thing to check is whether the flush operation has actually completed. The API is asynchronous, so it’s important to confirm the Azure-AsyncOperation (or Operation-Location) shows Succeeded.

That said, even after a successful flush, Azure metrics (including key count) can take a couple of minutes to update. So a delay there is normal.

The quickest way to confirm is to check at the Redis level,if DBSIZE returns 0, then the flush has worked and what you’re seeing is just metric lag. If it’s not 0, then the operation likely hasn’t completed or didn’t apply to the expected scope.

If you’re using geo-replication, you’ll also want to make sure the flush was done at the group level.

Happy to take a closer look if you can share what DBSIZE returns.
Pilladi Padma Sai Manisha 8,410 Reputation points Microsoft External Staff Moderator

2026-04-30T15:25:20.94+00:00

Hi
I hope you had a chance to review the information shared earlier, and I hope this information has been helpful! If you still have questions, please let us know what is needed in the comments so the question can be answered.

Answer 1

Hi Eddy Lynch,
Thankyou for reaching microsoft Q&A!
From what you’ve described, the flush call itself is working as expected in Azure Managed Redis, and the memory behavior you’re seeing is normal.

When you call the …/databases/default/flush API, it clears all keys from the database. However, Redis doesn’t immediately return freed memory back to the OS. Instead, it keeps that memory reserved for reuse, which is why the “used memory” metric can still show around ~2 GB even after a successful flush.

It’s also worth noting that the flush operation is asynchronous (returns a 202), so you should make sure the operation has fully completed by checking the Azure-AsyncOperation or Operation-Location endpoint. That said, even after it reports Succeeded, the memory metric may still look unchanged due to how Redis manages memory internally.

To confirm the flush actually worked, it’s better to validate at the data level. For example, running DBSIZE should return 0, INFO keyspace should show no keys, and any previously stored key should no longer be available. Those checks will confirm the cache is empty regardless of what the memory metric shows.

You might also see a small delay in Azure Monitor metrics updating, so it can help to wait a minute or two and refresh.

If your cache is part of an Active Geo-Replication setup, make sure the flush is performed at the group level as documented, otherwise the standard database flush should be sufficient.

If your goal is to actually reduce the reported memory usage, that typically only happens after a restart or scaling operation, since that’s when the allocator is reset.
When the flush happened, all existing keys would have been cleared, which explains the drop from ~8 to ~3. The current count likely reflects services restarting and gradually repopulating the cache with only the keys they immediately need.

This is generally expected for cache-backed workloads. You should see the key count increase over time as the application continues to rebuild its cache.

For reference: https://learn.microsoft.com/azure/azure-cache-for-redis/cache-administration#flush-data https://learn.microsoft.com/azure/azure-cache-for-redis/cache-best-practices-memory-management/ https://learn.microsoft.com/azure/azure-cache-for-redis/managed-redis/managed-redis-how-to-active-geo-replication#flush-operation

Let me know if you want help validating this on your cluster or checking anything specific.

Share via

Azure Managed Redis - REST API to flush

0 additional answers

Your answer