Apache Camel K

Apache Camel K is a lightweight integration platform, born on Kubernetes, with serverless superpowers.

Getting Started

Camel K allows to run integrations directly on a Kubernetes or OpenShift cluster. To use it, you need to be connected to a cloud environment or to a local cluster created for development purposes.

If you need help on how to create a local development environment based on Minishift or Minikube, you can follow the local cluster setup guide.

Installation

Make sure you apply specific configuration settings for your cluster before installing Camel K. Customized instructions are needed for the following cluster types:

Other cluster types (such as OpenShift clusters) should not need prior configuration.

To start using Camel K you need the "kamel" binary, that can be used to both configure the cluster and run integrations. Look into the release page for latest version of the kamel tool.

If you want to contribute, you can also build it from source! Refer to the contributing guide for information on how to do it.

Once you have the "kamel" binary, log into your cluster using the standard "oc" (OpenShift) or "kubectl" (Kubernetes) client tool and execute the following command to install Camel K:

kamel install

This will configure the cluster with the Camel K custom resource definitions and install the operator on the current namespace.

Important

Custom Resource Definitions (CRD) are cluster-wide objects and you need admin rights to install them. Fortunately this operation can be done once per cluster. So, if the kamel install operation fails, you’ll be asked to repeat it when logged as admin. For Minishift, this means executing oc login -u system:admin then kamel install --cluster-setup only for first-time installation.

Running an Integration

After the initial setup, you can run a Camel integration on the cluster by executing:

kamel run examples/Sample.java

During your first run you may see your console stuck with the following message:

integration "sample" created
integration "sample" in phase Waiting For Platform

Note	It will take some time for your integration to get started for the first time. This is because the camel-k operator has to pull and cache the camel-k builder images into the cluster’s registry.

You can follow this process by watching the pods in the namespace where the operator is running:

kubectl get pods -w

You should see something like:

$ kubectl get pods -w
NAME                                READY   STATUS      RESTARTS   AGE
camel-k-cache                       0/1     Completed   0          3m29s
camel-k-groovy-builder              1/1     Running     0          32s
camel-k-jvm                         0/1     Completed   0          61s
camel-k-jvm-builder                 0/1     Completed   0          2m13s
camel-k-operator-587b579567-fkz54   1/1     Running     1          43m
camel-k-groovy   0/1   Pending   0     0s
camel-k-groovy   0/1   ContainerCreating   0     0s
camel-k-groovy   1/1   Running   0     2s
camel-k-groovy   0/1   Completed   0     4s
camel-k-kotlin-builder   0/1   Pending   0     0s
camel-k-kotlin-builder   0/1   Pending   0     0s
camel-k-kotlin-builder   0/1   ContainerCreating   0     0s
camel-k-groovy-builder   0/1   Completed   0     75s
...

After a couple of minutes you can see your integration runnning. To see the output log of your integration container get the pod id:

kubectl get pods | grep sample
sample-67699569c4-724dv             1/1     Running     0          15m

And then see the logs using:

kubectl logs -f sample-67699569c4-724dv

An output like this should appear:

Starting the Java application using /opt/run-java/run-java.sh ...
...
2019-05-08 19:35:21.883 INFO  [main] DefaultCamelContext - Apache Camel 2.23.2 (CamelContext: camel-k) started in 0.874 seconds
2019-05-08 19:35:22.889 INFO  [Camel (camel-k) thread #2 - timer://tick] route1 - Hello Camel K!
2019-05-08 19:36:22.881 INFO  [Camel (camel-k) thread #2 - timer://tick] route1 - Hello Camel K!
...

A "Sample.java" file is included in the /examples folder of this repository. You can change the content of the file and execute the command again to see the changes.

Configure Integration properties

Properties associated to an integration can be configured either using a ConfigMap/Secret or by setting using the "--property" flag, i.e.

kamel run --property my.message=test examples/props.js

kamel run --configmap=<your name here> examples/props.js

Note: to create the config map first create a file called application.properties which contains lines with key=value pairs e.g. my.message="The text to display" Create the config map in the usual manner e.g.

kubectl create configmap <your name here>  --from-file=application.properties

Configure Integration Logging

camel-k runtime uses log4j2 as logging framework and can be configured through integration properties. If you need to change the logging level of various loggers, you can do so by using the logging.level prefix:

logging.level.org.apache.camel = DEBUG

Configure Integration Components

camel-k component can be configured programmatically inside an integration or using properties with the following syntax.

camel.component.${scheme}.${property} = ${value}

As example if you want to change the queue size of the seda component, you can use the following property:

camel.component.seda.queueSize = 10

Configure Maven settings

Maven settings can be configured by providing a custom settings.xml through a ConfigMap or secret. The source for maven settings can be configured using either the kamel CLI at installation time or by editing Camel K’s IntegrationPlatform:

cli:

Syntax

kamel install --maven-settings=configmap|secret:name[/key]

Example

kamel cmd --maven-settings=configmap:maven-settings/settings.xml

IntegrationPlatform:

apiVersion: camel.apache.org/v1
kind: IntegrationPlatform
metadata:
  labels:
    app: camel-k
  name: camel-k
spec:
  build:
    maven:
      settings:
        configMapKeyRef:
          key: settings.xml
          name: maven-settings

To quickly create a config map you can use the following command:

kubectl create configmap maven-settings --from-file=settings.xml

As an installation shortcut, additional maven repositories can be alternatively added using the --maven-repository option

kamel cmd --maven-repository http://repo1.my-company.com --maven-repository http://repo2.my-company.com

A repository url follow conventions used to configure additional repositories in Apache Karaf so it can be appended with zero or more of the following flags:

@snapshots: the repository contains snapshots
@noreleases: the repository does not contain any released artifacts
@id=repository.id: the id for the repository

kamel install --maven-repository http://repository.apache.org/content/groups/snapshots-group@id=apache@snapshots@noreleases

This results in:

<repositories>
  <repository>
    <id>apache</id>
    <url>http://repository.apache.org/content/groups/snapshots-group</url>
    <snapshots>
      <enabled>true</enabled>
    </snapshots>
    <releases>
      <enabled>false</enabled>
    </releases>
  </repository>
</repositories>

IMPORTANT

the options --maven-settings and --maven-repository are mutually exclusive

Configure Integration Volumes

It’s possible to mount persistent volumes into integration containers by using the -v or --volume flag. The format of volume flag value is similar to that of the docker CLI. But instead of specifying a host path to mount from, you reference the name of a PersistentVolumeClaim that you have already configured within the cluster. E.g

kamel run examples/Sample.java -v myPvcName:/some/path

Configure Environment Variables

It’s possible to configure environment variables for integration containers by using the -e or --env flag:

kamel run examples/Sample.java -e MY_ENV_VAR=some-value

Running Integrations in "Dev" Mode for Fast Feedback

If you want to iterate quickly on an integration to have fast feedback on the code you’re writing, you can use by running it in "dev" mode:

kamel run examples/Sample.java --dev

The --dev flag deploys immediately the integration and shows the integration logs in the console. You can then change the code and see the changes automatically applied (instantly) to the remote integration pod.

The console follows automatically all redeploys of the integration.

Here’s an example of the output:

[nferraro@localhost camel-k]$ kamel run examples/Sample.java --dev
integration "sample" created
integration "sample" in phase Building
integration "sample" in phase Deploying
integration "sample" in phase Running
[1] Monitoring pod sample-776db787c4-zjhfr[1] Starting the Java application using /opt/run-java/run-java.sh ...
[1] exec java -javaagent:/opt/prometheus/jmx_prometheus_javaagent.jar=9779:/opt/prometheus/prometheus-config.yml -XX:+UseParallelGC -XX:GCTimeRatio=4 -XX:AdaptiveSizePolicyWeight=90 -XX:MinHeapFreeRatio=20 -XX:MaxHeapFreeRatio=40 -XX:+ExitOnOutOfMemoryError -cp .:/deployments/* org.apache.camel.k.jvm.Application
[1] [INFO ] 2018-09-20 21:24:35.953 [main] Application - Routes: file:/etc/camel/conf/Sample.java
[1] [INFO ] 2018-09-20 21:24:35.955 [main] Application - Language: java
[1] [INFO ] 2018-09-20 21:24:35.956 [main] Application - Locations: file:/etc/camel/conf/application.properties
[1] [INFO ] 2018-09-20 21:24:36.506 [main] DefaultCamelContext - Apache Camel 2.22.1 (CamelContext: camel-1) is starting
[1] [INFO ] 2018-09-20 21:24:36.578 [main] ManagedManagementStrategy - JMX is enabled
[1] [INFO ] 2018-09-20 21:24:36.680 [main] DefaultTypeConverter - Type converters loaded (core: 195, classpath: 0)
[1] [INFO ] 2018-09-20 21:24:36.777 [main] DefaultCamelContext - StreamCaching is not in use. If using streams then its recommended to enable stream caching. See more details at http://camel.apache.org/stream-caching.html
[1] [INFO ] 2018-09-20 21:24:36.817 [main] DefaultCamelContext - Route: route1 started and consuming from: timer://tick
[1] [INFO ] 2018-09-20 21:24:36.818 [main] DefaultCamelContext - Total 1 routes, of which 1 are started
[1] [INFO ] 2018-09-20 21:24:36.820 [main] DefaultCamelContext - Apache Camel 2.22.1 (CamelContext: camel-1) started in 0.314 seconds

Dependencies and Component Resolution

Camel components used in an integration are automatically resolved. For example, take the following integration:

from("imap://admin@myserver.com")
  .to("seda:output")

Since the integration is using the "imap:" prefix, Camel K is able to automatically add the "camel-mail" component to the list of required dependencies. This will be transparent to the user, that will just see the integration running.

Automatic resolution is also a nice feature in --dev mode, because you are allowed to add all components you need without exiting the dev loop.

You can also use the -d flag to pass additional explicit dependencies to the Camel client tool:

kamel run -d mvn:com.google.guava:guava:26.0-jre -d camel-mina2 Integration.java

Not Just Java

Camel K supports multiple languages for writing integrations:

Table 1. Languages

Language	Description
Java	Integrations written in plain Java DSL are supported.
XML	Integrations written in plain XML DSL are supported (Spring XML or Blueprint not supported).
Groovy	Groovy `.groovy` files are supported (experimental).
JavaScript	JavaScript `.js` files are supported (experimental).
Kotlin	Kotlin Script `.kts` files are supported (experimental).

More information about supported languages is provided in the languages guide.

Integrations written in different languages are provided in the examples directory.

An example of integration written in JavaScript is the /examples/dns.js integration. Here’s the content:

// Lookup every second the 'www.google.com' domain name and log the output
from('timer:dns?period=1000')
    .routeId('dns')
    .setHeader('dns.domain')
        .constant('www.google.com')
    .to('dns:ip')
    .to('log:dns');

To run it, you need just to execute:

kamel run examples/dns.js

Traits

The details of how the integration is mapped into Kubernetes resources can be customized using traits. More information is provided in the traits section.

Monitoring the Status

Camel K integrations follow a lifecycle composed of several steps before getting into the Running state. You can check the status of all integrations by executing the following command:

kamel get

Contributing

We love contributions and we want to make Camel K great!

Contributing is easy, just take a look at our developer’s guide.

Uninstalling

If you really need to, it is possible to completely uninstall Camel K from OpenShift or Kubernetes with the following command, using the "oc" or "kubectl" tool:

# kubectl on plain Kubernetes
oc delete all,pvc,configmap,rolebindings,clusterrolebindings,secrets,sa,roles,clusterroles,crd -l 'app=camel-k'

Licensing

This software is licensed under the terms you may find in the file named LICENSE in this directory.

rareddy/camel-k