Camel K Operator fine tuning
Camel K Operators offers several possibility of customization. The default installation could be good in the most of the cases, but, we have a series of configuration that can be applied when you want to fine tune your Camel K operator and get the very best of it.
Some of the configuration can be defined after the operator is installed using the proper IntegrationPlatform specifications. Other fine tunings require to change the operator Deployment, so they either need to be provided before the installation procedure start, or after, with the subsequent operator restart.
| changing some of the default configuration may affect the behavior of your Camel K Operator. Make sure to understand how to properly tune each configuration. |
Most of these settings will require changing either an environment variable or other configuration of the operator Deployment. You need to verify how to apply those changes according the installation methodology chosen.
Installation name
You can install one or more Camel K operators (which must share CRDs). In such case you need to specify a unique operator name which will be used by each resource to know who is in charge to reconcile it. This behavior is controlled by the OPERATOR_ID environment variable (default value, camel-k).
Watch namespaces
Camel K operator can work in global (descoped) mode by watching all namespaces or namespaced (scoped). This behavior is controlled by the WATCH_NAMESPACE environment variable. If set to an empty value, then the operator will watch for resources in all namespaces. If it set to a given namespace, then it will only watch for resources in such a namespace.
Resource management
We provide certain configuration to better "operationalize" the Camel K Operator. More detailed information on the resource management page.
Monitoring
Camel K Operator provides certain parameters that would let you enable monitoring capabilities.
There is a default liveness service that is available at the operator Pod /healtz endpoint on port 8081. You can change the port parameter if you need to expose under a different port making sure to start the operator with the --health-port to the same value. You also need to change the Deployment template spec to reflect the change on the operator container liveness check.
There is a default metrics service as well available at the operator Pod /metrics endpoint on port 8080. The port can be also changed. In this case you only need to change the Deployment container operator, changing the --monitoring-port value to the port you want the operator to expose.
The service exposed on the /metrics endpoint is compatible with Prometheus. You can learn more about how to monitor Camel K Operator.
Logging
By default, the operator Pod logging level is set to INFO. You may require to change this value to log also DEBUG. In this case you need to provide an environment variable to the operator Deployment, LOG_LEVEL which can be set to debug. You can also lower the number of logging tracing by setting this value to error (advisable for production environments).
Environment Variables
The following environment variables can be configured on the operator Deployment to tune the operator behavior:
| Variable | Default | Description |
|---|---|---|
| empty | The namespace(s) the operator watches for Integration resources. When not set or empty, the operator watches all namespaces. |
| auto-detected | The namespace where the operator pod is running. |
| auto-detected | The name of the operator pod. |
|
| The operator ID used to identify which operator instance manages specific resources. Can also be set via |
|
| The base image used for building Integration images. Can also be set via |
|
| When set to |
|
| The log level for the operator. Valid values: |
|
| Directory where Kamelet catalog files are extracted during installation. |