Skip to main content
Version: 1.2

Configuration

Botkube backend allows you to specify source, executor, communication, and general Botkube settings. Check the related documents for more detailed explanation.

The configuration settings are read from two sources:

  • the configuration files specified by the BOTKUBE_CONFIG_PATHS environment variable or --config/-c flag. For example:

    export BOTKUBE_CONFIG_PATHS="global.yaml,team-b-specific.yaml"
    # or
    ./botkube --config "global.yaml,team-b-specific.yaml"

    You can split individual settings into multiple configuration files. The priority will be given to the last (right-most) file specified. Files with _ name prefix are always read as the last ones. See the merging strategy section for more details.

    note

    For Helm installation, Botkube uses _runtime_state.yaml and _startup_state.yaml files to store its internal state. Remember to keep these files in the BOTKUBE_CONFIG_PATHS environment variable.

  • the exported environment variables that overrides the configuration specified in the files.

Helm install options​

Advanced Helm install options are documented here.

Updating the configuration​

To update Botkube configuration, you can either:

  • upgrade Botkube installation with Helm,
  • or use dedicated @Botkube commands, to e.g. toggle notifications or edit Source Bindings. See the Usage document for more details.

If you wish to change the configuration with Helm, create a /tmp/values.yaml file that contains the new values and run:

botkube install -f /tmp/values.yaml

As both Helm release upgrade and some of the @Botkube commands modify the same configuration, it is merged during command execution. Whenever you specify a new value in the /tmp/values.yaml file, it will override the existing value in the configuration.

Preventing overrides by default Helm chart values​

Keep in mind that even if you don't specify custom values in the /tmp/values.yaml file, Helm can override the existing values with the default ones.

Consider the following config:

communications:
"default-group":
socketSlack:
enabled: true
botToken: "{botToken}"
appToken: "{appToken}"
channels:
"default":
name: general
notification:
disabled: false # default from the Helm chart
bindings:
sources:
- k8s-all-events # default from the Helm chart
# (...)

Assume that users ran the following commands:

@Botkube edit SourceBindings k8s-err-events, k8s-recommendation-events
@Botkube disable notifications

Which effectively result in the following config that Botkube sees:

communications:
"default-group":
socketSlack:
enabled: true
botToken: "{botToken}"
appToken: "{appToken}"
channels:
"default":
name: general
notification:
disabled: true # set by user command
bindings:
sources:
- k8s-err-events # set by user command
- k8s-recommendation-events # set by user command
# (...)

To persist the configuration that users provided, and not overwrite notification and source bindings values, run Helm upgrade with:

communications:
"default-group":
socketSlack:
channels:
"default":
name: general
notification: null # explicitly not use defaults from Helm chart
bindings:
sources: null # explicitly not use defaults from Helm chart
# (...) other values

The following properties need such null value during upgrade, if you want to keep the previous configuration:

  • communications.default-group.{communication-platform}.channels.default.notifications, where {communication-platform} is any communication platform supported except Microsoft Teams,
  • communications.default-group.{communication-platform}.channels.default.bindings.sources, where {communication-platform} is any communication platform supported except Microsoft Teams,
  • communications.default-group.teams.bindings.sources.

To learn more, read the Deleting a default key paragraph in Helm documentation.

Environment variables​

The individual communication settings can be specified via environment variables. They take priority and override the configuration specified in the file.

To construct the environment variable name, take any property from the configuration file and make it uppercase. Use the underscore for properties that are nested. Use the double underscore for all camelCase properties. Finally, add the BOTKUBE_ prefix.

For example, such configuration property from YAML:

settings:
kubectl:
defaultNamespace: "NAMESPACE"

is mapped to the BOTKUBE_SETTINGS_KUBECTL_DEFAULT__NAMESPACE environment variable.

This is a useful feature that allows you to store the overall configuration in a file, where sensitive data, such as tokens, can be put in environment variables. See the Tokens from Vault via CSI driver tutorial for an example use-case.

Merging strategy​

Botkube allows you to split individual settings into multiple configuration files. The following rules apply:

  • The priority will be given to the last (right-most) file specified.

  • Files with _ name prefix are always read as the last ones following the initial order.

  • Objects are merged together and primitive fields are overridden. For example:

    # a.yaml - first file
    settings:
    clusterName: dev-cluster
    configWatcher: true
    kubectl:
    enabled: false
    # _a.yaml - second file with `_` prefix
    settings:
    clusterName: demo-cluster
    # b.yaml - third file
    settings:
    kubectl:
    enabled: true
    # result
    settings:
    clusterName: demo-cluster
    configWatcher: true
    kubectl:
    enabled: true
  • The arrays items are not merged, they are overridden. For example:

    # a.yaml
    settings:
    kubectl:
    enabled: true
    commands:
    verbs:
    ["api-resources", "api-versions", "cluster-info", "describe", "diff", "explain", "get", "logs", "top", "auth"]
    # b.yaml
    settings:
    kubectl:
    commands:
    verbs: ["get", "logs", "top", "auth"]
    # result
    settings:
    kubectl:
    enabled: true
    commands:
    verbs: ["get", "logs", "top", "auth"]