Basic diagnostics

Here you can find the list of basic diagnostic actions that may help you look for bug causes.

In order to run the suggested commands, make sure that you have installed:

• helm
• kubectl

Agent

This section describes Botkube agent related diagnostic.

Agent version

The easiest way to check the Agent version is to get the Docker image:

kubectl get deploy botkube -n botkube -o=jsonpath="{'Used images\n'}{range .spec.template.spec.containers[*]}{.name}{':\t'}{.image}{'\n'}{end}"

You should get an output similar to this:

Used images
botkube:    ghcr.io/kubeshop/botkube:v1.4.0
cfg-watcher:    ghcr.io/kubeshop/k8s-sidecar:in-cluster-config

The botkube is the agent image. The container image tag (v1.4.0) is the version in which it was deployed on the cluster.

Agent health

To check if the Agent Pods are in the Running state, run:

kubectl get pod -n botkube -l app=botkube

All the containers from Pods should be in the Running status. Restarts' number higher than one may also indicate problems, e.g. not enough resource, lack of permissions, network timeouts, etc.

Agent logs

If the Botkube Agent is healthy, you should be able to track any bug by checking the logs. To check the logs, run:

kubectl logs -n botkube -l app=botkube -c botkube

To get all logs specify --tail=-1, otherwise only 10 last lines are displayed.

To check the logs since a given time, use the --since-time or --since flag, for example:

--since-time=2020-03-30T10:02:08Z

--since=3h

Agent configuration

Select a tab to use a tool of your choice for getting Botkube configuration:

Botkube CLI

note

The botkube config get command is available from the v1.4.0 version.

Install Botkube CLI and run:

botkube config get > /tmp/bk-config.yaml

kubectl

1. Download export-config-job-tpl definition into /tmp/export-config-job-tpl.
2. Make sure that the previous job is deleted:

   kubectl delete job botkube-migration -n botkube
3. Run a new export config Job based on the configuration from installed Botkube deployment:

   kubectl get deploy botkube -n botkube -o=go-template-file=/tmp/export-config-job-tpl | kubectl create -f -
4. Get configuration data:

   kubectl get cm botkube-config-exporter -ojsonpath='{ .data.config\.yaml }'

Agent restart

When Pods are unhealthy, or if the operation processing is stuck, you can restart the Pod using this command:

kubectl delete po -n botkube -l app=botkube

Agent debug logging

In order to change the logging level to debug, run:

helm upgrade botkube botkube/botkube -n botkube --set settings.log.level="debug" --reuse-values

If the Botkube agent Pod isn't restarted automatically, restart it manually.

Check configured plugin repositories

Select a tab to use a tool of your choice for checking plugin repository configuration:

yq

Install yq and run:

helm get values botkube --all -oyaml | yq '.plugins'

jq

Install jq and run:

helm get values botkube --all -ojson | jq '{ plugins }'

grep

helm get values botkube --all -oyaml | grep -A 10 'plugins:'

The output may contain more occurrence of a plugin section, so you need to fine the one that has the repositories defined.