Description:
I have deployed PMM on kubernetes cluster using helm chart and also using Nginx as an ingress controller.
I see this kind of error every now and then in Nginx logs
2025/03/14 02:10:00 [error] 21683#21683: *66801000 client intended to send too large chunked body: 104868960 bytes while sending request to upstream, client: <kubernetes-nginx-pod-ip>, server: mypmm.com, request: "POST /agent.Agent/Connect HTTP/2.0", upstream: "grpcs://<kubernetes-nginx-pod-ip>:443", host: "mypmm.com"
at first, I didn’t specified nginx.kubernetes.io/proxy-body-size so I thought this could solve this problem, after I increased nginx.kubernetes.io/proxy-body-size to 3m(by that time the bytes in the error log was 1.04mb) then log change to 3.04mb is too large chunked body.
So I increased it again to 100m then I see current log which is client intended to send too large chunked body: 104868960 bytes while sending request to upstream
Steps to Reproduce:
Deploy PMM helm chart on the kubernetes that using Nginx as an ingress controller
Version:
helm chart: 1.3.17
application: 2.43.1
Logs:
Nginx pod:
2025/03/14 02:10:00 [error] 21683#21683: *66801000 client intended to send too large chunked body: 104868960 bytes while sending request to upstream, client: <kubernetes-nginx-pod-ip>, server: mypmm.com, request: "POST /agent.Agent/Connect HTTP/2.0", upstream: "grpcs://<kubernetes-nginx-pod-ip>:443", host: "mypmm.com"
Expected Result:
sending request without chunk size error
Actual Result:
chunk size always exceeds the limit despite increasing nginx.kuberenetes.io/proxy-body-size