Queries got queued after disk size/performance increase on k8s (AKS)

Help Workload Profiling
1

We have increased the disk size for our SingleStoreDB leaves from 4TB to 5TB (azure premium SSD P50 to P60).

Cluster is running on AKS with nodepool based on the Standard_E20s_v5 VM.

After that the queries got queued and the execution time degraded from below second to up to 10s.

We are at version 8.9.12 ,1 master, two children, 2 leaves, height 1 (8vCP/64GB).

the Azure monitoring does not show any capping at disks/VMs occurrences, yet the SingleStore Studio also does not report any high usage of CPU/RAM.

Are there any guidance we can follow to check why suddenly queries got queued and if it was related to that disk change ?

Or maybe there is fix for that in some next versions ?

2

I don’t have a direct answer for you. Consider upgrading to version 9.0 because it includes improvements to workload management and resource handling. If you haven’t already, consider going to SingleStore | Develop with SQrL, Your SingleStore Copilot. and enter just what you typed in your question. It will point out several things to consider regarding workload management, plan cache size limits that are a function of disk size, and so on.

3

it seems we’ve been struggled by IOPS threshold.

after VM/disk upgrade to higher SKU the issue gone