A Portal to a Portal: Kubernetes - tailing Pod logs

Wednesday, 13 October 2021

Kubernetes - tailing Pod logs - TIL

I saw this on Twitter a few days back: -

but hadn't yet got around to playing with it.

I thought I'd test it with a brand new K8s 1.20 cluster running on IBM Kubernetes Service (IKS).

Firstly, I queried the pods that were running: -

kubectl get pods -A

NAMESPACE NAME READY STATUS RESTARTS AGE

ibm-system addon-catalog-source-2x7hj 1/1 Running 0 41h
ibm-system catalog-operator-578f7c8857-666wd 1/1 Running 0 41h
ibm-system olm-operator-6c45d79d96-pjtmr 1/1 Running 0 41h
kube-system calico-kube-controllers-78ccd56cd7-wqgtf 1/1 Running 0 41h
kube-system calico-node-pg6vv 1/1 Running 0 41h
kube-system calico-typha-ddd44968b-86cgs 1/1 Running 0 41h
kube-system calico-typha-ddd44968b-ffxmt 0/1 Pending 0 41h
kube-system calico-typha-ddd44968b-mqjrb 0/1 Pending 0 41h
kube-system coredns-7fc9f85d9c-5rwwv 1/1 Running 0 41h
kube-system coredns-7fc9f85d9c-bxtts 1/1 Running 0 41h
kube-system coredns-7fc9f85d9c-qk6gv 1/1 Running 0 41h
kube-system coredns-autoscaler-9cccfb98d-mw9qj 1/1 Running 0 41h
kube-system dashboard-metrics-scraper-7c75dcd466-d5b9f 1/1 Running 0 41h
kube-system ibm-keepalived-watcher-856k6 1/1 Running 0 41h
kube-system ibm-master-proxy-static-10.144.213.225 2/2 Running 0 41h
kube-system kubernetes-dashboard-659cd5b798-thd57 1/1 Running 0 41h
kube-system metrics-server-b7bc76594-4fdg2 2/2 Running 0 41h
kube-system vpn-546847fcbf-dzzml 1/1 Running 0 41h

and chose the metrics-server-b7bc76594-4fdg2 pod, because it's running two containers: -

kubectl get pod metrics-server-b7bc76594-4fdg2 --namespace kube-system --output json | jq -r .spec.containers[].name

metrics-server

metrics-server-nanny

Grabbing the labels for this pod: -

kubectl get pod metrics-server-b7bc76594-4fdg2 --namespace kube-system --output json | jq .metadata.labels

{

"cert-checksum": "6457ed123878693e37fbde8b4a0abf966ae050c3",

"k8s-app": "metrics-server",

"pod-template-hash": "b7bc76594",

"version": "v0.4.4"

}

and choosing the k8s-app label, I followed Lili's advice: -

kubectl logs --namespace kube-system --selector k8s-app=metrics-server --container metrics-server --tail=-1

I1011 17:10:21.606350 1 dynamic_serving_content.go:130] Starting serving-cert::/etc/metrics-server-certs/tls.crt::/etc/metrics-server-certs/tls.key

I1011 17:10:21.606421 1 configmap_cafile_content.go:202] Starting client-ca::kube-system::extension-apiserver-authentication::client-ca-file

I1011 17:10:21.626860 1 shared_informer.go:240] Waiting for caches to sync for client-ca::kube-system::extension-apiserver-authentication::client-ca-file

I1011 17:10:21.646520 1 secure_serving.go:197] Serving securely on [::]:4443

I1011 17:10:21.627139 1 requestheader_controller.go:169] Starting RequestHeaderAuthRequestController

I1011 17:10:21.647378 1 shared_informer.go:240] Waiting for caches to sync for RequestHeaderAuthRequestController

I1011 17:10:21.647922 1 tlsconfig.go:240] Starting DynamicServingCertificateController

I1011 17:10:21.606461 1 configmap_cafile_content.go:202] Starting client-ca::kube-system::extension-apiserver-authentication::requestheader-client-ca-file

I1011 17:10:21.686383 1 shared_informer.go:240] Waiting for caches to sync for client-ca::kube-system::extension-apiserver-authentication::requestheader-client-ca-file

I1011 17:10:22.007270 1 shared_informer.go:247] Caches are synced for RequestHeaderAuthRequestController

I1011 17:10:22.046837 1 shared_informer.go:247] Caches are synced for client-ca::kube-system::extension-apiserver-authentication::client-ca-file

I1011 17:10:22.126237 1 shared_informer.go:247] Caches are synced for client-ca::kube-system::extension-apiserver-authentication::requestheader-client-ca-file

and: -

kubectl logs --namespace kube-system --selector k8s-app=metrics-server --container metrics-server-nanny --tail=-1

ERROR: logging before flag.Parse: I1011 17:10:03.951537 1 pod_nanny.go:68] Invoked by [/pod_nanny --config-dir=/etc/config --cpu=40m --extra-cpu=0.5m --memory=40Mi --extra-memory=4Mi --threshold=5 --deployment=metrics-server --container=metrics-server --poll-period=300000 --estimator=exponential --use-metrics=true]

ERROR: logging before flag.Parse: I1011 17:10:03.951645 1 pod_nanny.go:69] Version: 1.8.12

ERROR: logging before flag.Parse: I1011 17:10:03.951673 1 pod_nanny.go:85] Watching namespace: kube-system, pod: metrics-server-b7bc76594-4fdg2, container: metrics-server.

ERROR: logging before flag.Parse: I1011 17:10:03.951684 1 pod_nanny.go:86] storage: MISSING, extra_storage: 0Gi

ERROR: logging before flag.Parse: I1011 17:10:03.968655 1 pod_nanny.go:116] cpu: 40m, extra_cpu: 0.5m, memory: 40Mi, extra_memory: 4Mi

ERROR: logging before flag.Parse: I1011 17:10:03.968697 1 pod_nanny.go:145] Resources: [{Base:{i:{value:40 scale:-3} d:{Dec:<nil>} s:40m Format:DecimalSI} ExtraPerNode:{i:{value:5 scale:-4} d:{Dec:<nil>} s: Format:DecimalSI} Name:cpu} {Base:{i:{value:41943040 scale:0} d:{Dec:<nil>} s: Format:BinarySI} ExtraPerNode:{i:{value:4194304 scale:0} d:{Dec:<nil>} s:4Mi Format:BinarySI} Name:memory}]

In other words, I used the --selector switch to grab logs for the two containers by their label, rather than knowing/caring about the pod name.

I could've done much the same BUT via a much more complicated route using grep and awk : -

kubectl logs --namespace kube-system $(kubectl get pods --namespace kube-system | grep metrics-server | awk '{print $1}') -c metrics-server