Kubernetes tail. Streams logs from all containers of all matched pods. Match pods by service, replicaset, deployment, and others. Adjusts to a changing cluster - pods are added and removed from logging as they fall in or out of the selection.
With no arguments, kail matches all pods in the cluster. You can control the matching pods with arguments which select pods based on various criteria.
All flags can be set by an environment variable - KAIL_POD=foo kail
is the same as kail --pod foo
.
Flag | Selection |
---|---|
-l, --label LABEL-SELECTOR |
match pods based on a standard label selector |
-p, --pod NAME |
match pods by name |
-n, --ns NAMESPACE-NAME |
match pods in the given namespace |
--svc NAME |
match pods belonging to the given service |
--rc NAME |
match pods belonging to the given replication controller |
--rs NAME |
match pods belonging to the given replica set |
-d, --deploy NAME |
match pods belonging to the given deployment |
--sts NAME |
match pods belonging to the given statefulset |
-j, --job NAME |
match pods belonging to the given job |
--node NODE-NAME |
match pods running on the given node |
--ing NAME |
match pods belonging to services targeted by the given ingress |
-c, --containers CONTAINER-NAME |
restrict which containers logs are shown for |
--ignore LABEL-SELECTOR |
Ignore pods that the selector matches. (default: kail.ignore=true ) |
--current-ns |
Match pods in the namespace specified in Kubernetes' "current context" |
--ignore-ns NAME |
Ignore pods in the given namespaces. Overridden by --ns , --current-ns . (default: kube-system ) |
When selecting objects by NAME
(--svc
, --pod
, etc...), you can either qualify it with a namespace to restrict the selection to the given namespace, or select across all namespaces by giving just the object name.
Example:
# match pods belonging to a replicaset named 'workers' in any namespace.
$ kail --rs workers
# match pods belonging to the replicaset named 'workers' only in the 'staging' namespace
$ kail --rs staging/workers
If the same flag is used more than once, the selectors for that flag are "OR"ed together.
# match pods belonging to a replicaset named "workers" or "db"
$ kail --rs workers --rs db
Different flags are "AND"ed together:
# match pods belonging to both the service "frontend" and the deployment "webapp"
$ kail --svc frontend --deploy webapp
Flag | Description |
---|---|
-h, --help |
Display help and usage |
--context CONTEXT-NAME |
Use the given Kubernetes context |
--dry-run |
Print initial matched pods and exit |
--log-level LEVEL |
Set the logging level (default: error ) |
--log-file PATH |
Write output to PATH (default: /dev/stderr ) |
--since DURATION |
Display logs as old as given duration. Ex: 5s , 2m , 1.5h or 2h45m (defaults: 1s ). See here for more information on the duration format. |
-o, --output |
You can choose to display logs in default, raw (without prefix), json, pretty json and zerolog formats. |
$ brew tap boz/repo
$ brew install boz/repo/kail
$ kubectl krew install tail
$ kubectl tail -h
Kail binaries for Linux and OSX can be found on the latest release page. Download and install into your $GOPATH/bin
.
kubectl
The docker image abozanich/kail is available for running kail
from within a kubernetes pod via kubectl
.
Note: be sure to include the kail.ignore=true
label, otherwise... it's logging all the way down.
Example:
# match all pods - synonymous with 'kail' from the command line
$ kubectl run -it --rm -l kail.ignore=true --restart=Never --image=abozanich/kail kail
# match pods belonging to service 'api' in any namespace - synonymous with 'kail --svc api'
$ kubectl run -it --rm -l kail.ignore=true --restart=Never --image=abozanich/kail kail -- --svc api
$ go get -d github.com/boz/kail
$ cd $GOPATH/src/github.com/boz/kail
$ make install-deps
$ make
$ minikube start
$ ./_example/demo.sh start
$ ./kail
# install image into minikube and run via kubectl
$ make image-minikube
$ kubectl run -it --rm -l kail.ignore=true --restart=Never --image=kail kail
Start minikube with a fair amount of CPU and memory.
$ minikube start --cpus 4 --memory 8192
Start stress.yml
resources.
$ kubectl create -f _example/stress.yml
Wait a while for the pods to run
$ kubectl get pods --namespace stress | grep Running | wc -l
100
Run kail
./kail --ns stress