Setting Up TeamCity for Kubernetes
Thanks to the Kubernetes Support plugin, TeamCity can run build agents in your Kubernetes cluster.
note
If you were using the Helm build runner, included in the external Kubernetes Support plugin, note that the built-in integration does not comprise the Helm runner. Refer to our upgrade notes for more details.
TeamCity integration with Kubernetes does not depend on the kubectl tool and thus does not require installing it in your cluster.
Make sure the TeamCity user is allowed to perform writing operations in the Kubernetes namespace used by TeamCity agents.
You might also require to configure the following privileges for your Kubernetes user role:
Pods:
get,create,list,delete.Deployments:
list,get— if you want to create an agent pod using a deployment configuration.Namespaces:
get,list— to allow TeamCity to suggest the namespaces available on your server.
Here is an example of setting up all required permissions via Kubernetes RBAC:
apiVersion: rbac.authorization.k8s.io/v1
kind: Role
metadata:
name: teamcity:manage-agents
rules:
- apiGroups: [""]
resources: ["namespaces"]
verbs: ["list", "get"]
- apiGroups: [""]
resources: ["pods"]
verbs: ["get", "create", "list", "delete"]
- apiGroups: ["extensions", "apps"]
resources: ["deployments"]
verbs: ["list", "get"]
---
apiVersion: rbac.authorization.k8s.io/v1
kind: RoleBinding
metadata:
name: teamcity:manage-agents
roleRef:
apiGroup: rbac.authorization.k8s.io
kind: Role
name: teamcity:manage-agents
subjects:
# proper RoleBinding subject depends on your Authentication strategy
# use one of examples below
# if you use OIDC/Certificate auth strategies
- kind: User
name: teamcity
# if you use Service account
- kind: ServiceAccount
name: teamcityTo establish the integration with Kubernetes, you need to create a dedicated cloud profile in TeamCity. Open the settings of the required project and, under the Cloud Profiles section, click Create new profile.
The table below lists options specific for Kubernetes cloud profiles.
Option | Description |
|---|---|
Kubernetes API server URL | Specify the URL of the Kubernetes API server. |
Certificate Authority (CA) | Enter the content of the CA certificate for your cluster. |
Kubernetes namespace | Specify a required Kubernetes namespace. Leave empty to use the default namespace. |
Authentication strategy | Select the required authentication strategy. Depending on the selected strategy, the set of extra options will vary. Refer to the Kubernetes documentation for details on available options.
|
After configuring the general Kubernetes settings, you can proceed with adding a new build agent image.
Click Add image and configure its options:
Option | Description |
|---|---|
Pod specification | Select one of the three types described below. |
Agent pool | Assign the launched instances to a specific agent pool. |
Agent name prefix | (Optional) Show the same prefix for all instances of this image, when displaying them in TeamCity. |
Max number of instances | (Optional) Set this number if you want to limit how many instances are launched based on this image. |
tip
To run a build agent in Kubernetes, TeamCity will create a dedicated pod in your cluster.
Select this option to run a new pod based on a specific deployment configuration created in your Kubernetes cluster. If you are using Kubernetes Dashboard, you can find the list of available deployments under Workloads | Deployments.
You can create a simple deployment from the latest TeamCity agent image as follows:
create deployment simpledep --image="jetbrains/teamcity-agent:latest"
deployment.apps/simpledep createdSelect this option to run a single container based on any given agent Docker image from Docker Hub. You'll need to specify:
Option | Description |
|---|---|
Docker image name | Name of the Docker image, similarly to the one used with the |
Image pull policy | The policy of updating the image. |
Command | (Optional) Set the command run by the container. |
Command arguments | (Optional) Set the command arguments. |
Enter your own pod template in a YAML format. Use this option when you need to specify additional configuration for a launched pod, such as resources' requests/limits or credentials. This option is alternative to the template from deployment specification and allows editing the configuration directly in the TeamCity UI/DSL.
Example of a simple template:
apiVersion: v1
kind: Pod
metadata:
labels:
app: teamcity-agent
spec:
restartPolicy: Never
containers:
- name: teamcity-agent
image: jetbrains/teamcity-agent
resources:
limits:
memory: "2Gi"To create and setup k8s profiles and images, create KubernetesCloudProfile and KubernetesCloudImage class instances under the project's features group. Set an image profileId property to the profile id value to assign this image to the target profile.
project {
// ...
features {
// ...
kubernetesCloudProfile {
id = "kube-1"
name = "K8S Agents"
description = "EKS"
serverURL = "https://myteamcityserver.com"
terminateIdleMinutes = 30
apiServerURL = "https://123.gr7.eu-west-1.eks.amazonaws.com"
caCertData = "abc"
authStrategy = eks {
accessId = "aws_access_id"
secretKey = "aws_key"
clusterName = "my-k8s-cluster"
}
}
kubernetesCloudImage {
id = "PROJECT_EXT_10"
profileId = "kube-1"
agentPoolId = "21"
agentNamePrefix = "k8s-singlec"
maxInstancesCount = 5
podSpecification = runContainer {
dockerImage = "jetbrains/teamcity-agent"
}
}
kubernetesCloudImage {
id = "PROJECT_EXT_11"
profileId = "kube-1"
agentPoolId = "21"
agentNamePrefix = "k8s-pspec"
maxInstancesCount = 5
podSpecification = customTemplate {
customPod = """
apiVersion: v1
kind: Pod
metadata:
labels:
app: teamcity-agent
spec:
restartPolicy: Never
containers:
- name: teamcity-agent
image: jetbrains/teamcity-agent
resources:
limits:
memory: "2Gi"
""".trimIndent()
}
}
}
}See KubernetesCloudProfile and KubernetesCloudImage for more examples and API descriptions.