This small Go app that just watches one folder for specific files and tail them on the standard output
Use Case: run it as a sidecar in a Kubernetes Pod, to watch for the main container's logs and stream them so you can access them with kubectl logs <your pod> -c <watchfiletail container name>
Here is an application that is logging some informations in a log file, here /var/log/logging-app/logging-app.log:
---
apiVersion: v1
kind: Pod
metadata:
name: logging-app-pod-host-log
spec:
containers:
- name: logging-app
image: alpine:latest
command: ["/bin/sh", "-c"]
args:
- |
echo "starting"
while true
do
echo $(date) >> /var/log/logging-app/logging-app.log
sleep 5
done
echo "done"
volumeMounts:
- name: logging-app-log-dir
mountPath: /var/log/logging-app
volumes:
- name: logging-app-log-dir
emptyDir: {}
terminationGracePeriodSeconds: 1When using kubectl logs logging-app-pod-host-log, or better, Stern, you only see a starting log message.
The logStreamer is here to tail -f the log files and print them on stdout so Kubernetes can grab them.
You can add the log streamer like so:
---
apiVersion: v1
kind: Pod
metadata:
name: logging-app-pod-host-log
spec:
containers:
- name: logging-app
image: alpine:latest
command: ["/bin/sh", "-c"]
args:
- |
echo "starting"
while true
do
echo $(date) >> /var/log/logging-app/logging-app.log
sleep 5
done
echo "done"
volumeMounts:
- name: logging-app-log-dir
mountPath: /var/log/logging-app
- name: logstreamer
image: ghcr.io/prune998/watchfiletail:v0.0.4
env:
- name: FOLDERPATH
value: "/var/log/logging-app"
- name: LOGLEVEL
value: info
- name: FILEMATCH
value: "^logging-app.*"
resources:
limits:
cpu: "100m"
memory: 128Mi
requests:
cpu: "100m"
memory: 128Mi
volumeMounts:
- name: logging-app-log-dir
mountPath: /var/log/logging-app
volumes:
- name: logging-app-log-dir
emptyDir: {}
terminationGracePeriodSeconds: 1Add a mecanisme to delete old log files...