AKS
AKS copied to clipboard
Azure
[BUG] AGIC deployment stuck in endless Eviction loop
Describe the bug The AGIC deployment manifest requests 20mi of memory, which is lower than typical usage. My node is experiencing memory pressure and the AGIC pod continues to get evicted and rescheduled on the same node within seconds.
To Reproduce Steps to reproduce the behavior:
- Load up the system node(s) such that they are each flagged with a MemoryPressure condition
- Watch as the AGIC pod continually gets evicted shortly after startup when memory consumption goes above 20mi
Expected behavior The AGIC pod should run without getting evicted.
Screenshots
Environment (please complete the following information):
- Kubernetes version 1.26.3
- AGIC version 1.5.3
Additional context I have scaled my node pool out to alleviate the memory pressure as a workaround. Ideally, the 20mi memory request size should be increased to better align with memory requirements of the pod.
@chasewilson, could you help take a look or add the networking PM to this issue?
![microsoft-github-policy-service[bot] avatar](https://avatars.githubusercontent.com/in/95686?v=4 "Published by a pub.dev verified publisher"){kind=small}
@sabbour, @JackStromberg would you be able to assist?