Containerized AG sizing or scaling

[Mizső]

Hi Folks,

At one of our clinets after the 1.315 AG upgrade some containerized AGs (3 out of 25) could not restart with memory shortage (ClassicFullstak). I have not found any information about the containerized AG resourse sizing or scaling. Those AGs had memory problem which ones served more the 40 workernodes.

Messages:

025-07-04 14:39:51 UTC INFO [<collector>] [<platform>, MemoryTrackerImpl] 90% memory usage. [Suppressing further messages for 1 minute]

2025-07-04 14:40:29 UTC WARNING [<collector>] [<platform>, MemoryTrackerImpl] Heap memory shortage detected: 99% memory usage in memory pool 'G1 Old Gen'

One of the clusters view without kubernets API information until the resource increase:

From the v1beta1 to v1beta5 the AG recommended sizing is:

 resources:

      limits:

        cpu: '1'

        memory: 1536Mi

      requests:

        cpu: 500m

        memory: 512Mi

We have increased it to (in order to remediate the issue): 

resources:

      limits:

        cpu: '1'

        memory: 2048Mi

      requests:

        cpu: 500m

        memory: 1024Mi

Is there any official recommendation for containerized AGs resources or scaling?

Thanks in advance.

Best regards,

János

[radek_jasinski · Accepted solution]

Hi @Mizső 

You can find it in point 4: https://docs.dynatrace.com/docs/shortlink/ag-container-deploy#deployment

 

[Mizső]

Hi @radek_jasinski,

Thank you very much!!!

Indeed yellow clusters were affected with more than 1000 pods. We were lucky till now with the default recommended 1,5 GB limit. 

Best regards,

János