»Kubernetes

For a step by step tutorial, visit HashiCorp Learn.

Builders

Kubernetes uses Docker images for building, which are generated by these builders:

kubernetes (platform)

Deploy the application into a Kubernetes cluster using Deployment objects.

Interface

Examples

use "kubernetes" {
    image_secret = "registry_secret"
    replicas = 3
    probe_path = "/_healthz"
}

Required Parameters

These parameters are used in the use stanza for this plugin.

pod (category)

The configuration for a pod.

Pod describes the configuration for a pod when deploying.

pod.container (category)

Container describes the commands and arguments for a container config.

pod.container.args

An array of string arguments to pass through to the container.

Type: list of string
Optional

pod.container.command

An array of strings to run for the container.

Type: list of string
Optional

pod.container.cpu (category)

Cpu resource configuration.

CPU lets you define resource limits and requests for a container in a deployment.

pod.container.cpu.limit

Maximum amount of cpu to give the container. Supports m to indicate milli-cores.

Type: string
Optional

pod.container.cpu.request

How much cpu to give the container in cpu cores. Supports m to indicate milli-cores.

Type: string
Optional

pod.container.memory (category)

Memory resource configuration.

Memory lets you define resource limits and requests for a container in a deployment.

pod.container.memory.limit

Maximum amount of memory to give the container. Supports k for kilobytes, m for megabytes, and g for gigabytes.

Type: string
Optional

pod.container.memory.request

How much memory to give the container in bytes. Supports k for kilobytes, m for megabytes, and g for gigabytes.

Type: string
Optional

pod.container.name

Name of the container.

Type: string
Optional

pod.container.port (category)

A port and options that the application is listening on.

Used to define and expose multiple ports that the application or process is listening on for the container in use. Can be specified multiple times for many ports.

pod.container.port.host_ip

What host IP to bind the external port to.

Type: string
Optional

pod.container.port.host_port

The corresponding worker node port.

Number of port to expose on the host. If specified, this must be a valid port number, 0 < x < 65536. If HostNetwork is specified, this must match ContainerPort. Most containers do not need this.

Type: uint
Optional

pod.container.port.name

Name of the port.

If specified, this must be an IANA_SVC_NAME and unique within the pod. Each named port in a pod must have a unique name. Name for the port that can be referred to by services.

Type: string

pod.container.port.port

The port number.

Number of port to expose on the pod's IP address. This must be a valid port number, 0 < x < 65536.

Type: uint

pod.container.port.protocol

Protocol for port. Must be UDP, TCP, or SCTP.

Type: string
Optional
Default: TCP

pod.container.probe (category)

Configuration to control liveness and readiness probes.

Probe describes a health check to be performed against a container to determine whether it is alive or ready to receive traffic.

pod.container.probe.failure_threshold

Number of times a liveness probe can fail before the container is killed.

FailureThreshold * TimeoutSeconds should be long enough to cover your worst case startup times.

Type: uint
Optional
Default: 5

pod.container.probe.initial_delay

Time in seconds to wait before performing the initial liveness and readiness probes.

Type: uint
Optional
Default: 5

pod.container.probe.timeout

Time in seconds before the probe fails.

Type: uint
Optional
Default: 5

pod.container.probe_path

The HTTP path to request to test that the application is running.

Without this, the test will simply be that the application has bound to the port.

Type: string
Optional

pod.container.resources

A map of resource limits and requests to apply to a container on deploy.

Resource limits and requests for a container. This exists to allow any possible resources. For cpu and memory, use those relevant settings instead. Keys must start with either limits_ or requests_. Any other options will be ignored.

Type: map of string to string
Optional

pod.container.static_environment

Environment variables to control broad modes of the application.

Environment variables that are meant to configure the container in a static way. This might be control an image that has multiple modes of operation, selected via environment variable. Most configuration should use the waypoint config commands.

Type: map of string to string
Optional

pod.security_context (category)

Holds pod-level security attributes and container settings.

pod.security_context.fs_group

A special supplemental group that applies to all containers in a pod.

Type: int64

pod.security_context.run_as_non_root

Indicates that the container must run as a non-root user.

Type: bool

pod.security_context.run_as_user

The UID to run the entrypoint of the container process.

Type: int64

pod.sidecar (category)

A sidecar container within the same pod.

Another container to run alongside the app container in the kubernetes pod. Can be specified multiple times for multiple sidecars.

pod.sidecar.container (category)

Container describes the commands and arguments for a container config.

pod.sidecar.container.args

An array of string arguments to pass through to the container.

Type: list of string
Optional

pod.sidecar.container.command

An array of strings to run for the container.

Type: list of string
Optional

pod.sidecar.container.cpu (category)

Cpu resource configuration.

CPU lets you define resource limits and requests for a container in a deployment.

pod.sidecar.container.cpu.limit

Maximum amount of cpu to give the container. Supports m to indicate milli-cores.

Type: string
Optional

pod.sidecar.container.cpu.request

How much cpu to give the container in cpu cores. Supports m to indicate milli-cores.

Type: string
Optional

pod.sidecar.container.memory (category)

Memory resource configuration.

Memory lets you define resource limits and requests for a container in a deployment.

pod.sidecar.container.memory.limit

Maximum amount of memory to give the container. Supports k for kilobytes, m for megabytes, and g for gigabytes.

Type: string
Optional

pod.sidecar.container.memory.request

How much memory to give the container in bytes. Supports k for kilobytes, m for megabytes, and g for gigabytes.

Type: string
Optional

pod.sidecar.container.name

Name of the container.

Type: string
Optional

pod.sidecar.container.port (category)

A port and options that the application is listening on.

Used to define and expose multiple ports that the application or process is listening on for the container in use. Can be specified multiple times for many ports.

pod.sidecar.container.port.host_ip

What host IP to bind the external port to.

Type: string
Optional

pod.sidecar.container.port.host_port

The corresponding worker node port.

Number of port to expose on the host. If specified, this must be a valid port number, 0 < x < 65536. If HostNetwork is specified, this must match ContainerPort. Most containers do not need this.

Type: uint
Optional

pod.sidecar.container.port.name

Name of the port.

If specified, this must be an IANA_SVC_NAME and unique within the pod. Each named port in a pod must have a unique name. Name for the port that can be referred to by services.

Type: string

pod.sidecar.container.port.port

The port number.

Number of port to expose on the pod's IP address. This must be a valid port number, 0 < x < 65536.

Type: uint

pod.sidecar.container.port.protocol

Protocol for port. Must be UDP, TCP, or SCTP.

Type: string
Optional
Default: TCP

pod.sidecar.container.probe (category)

Configuration to control liveness and readiness probes.

Probe describes a health check to be performed against a container to determine whether it is alive or ready to receive traffic.

pod.sidecar.container.probe.failure_threshold

Number of times a liveness probe can fail before the container is killed.

FailureThreshold * TimeoutSeconds should be long enough to cover your worst case startup times.

Type: uint
Optional
Default: 5

pod.sidecar.container.probe.initial_delay

Time in seconds to wait before performing the initial liveness and readiness probes.

Type: uint
Optional
Default: 5

pod.sidecar.container.probe.timeout

Time in seconds before the probe fails.

Type: uint
Optional
Default: 5

pod.sidecar.container.probe_path

The HTTP path to request to test that the application is running.

Without this, the test will simply be that the application has bound to the port.

Type: string
Optional

pod.sidecar.container.resources

A map of resource limits and requests to apply to a container on deploy.

Type: map of string to string
Optional

pod.sidecar.container.static_environment

Environment variables to control broad modes of the application.

Type: map of string to string
Optional

pod.sidecar.image

Image of the sidecar container.

Type: string

Optional Parameters

These parameters are used in the use stanza for this plugin.

annotations

Annotations to be added to the application pod.

Annotations are added to the pod spec of the deployed application. This is useful when using mutating webhook admission controllers to further process pod events.

Type: map of string to string
Optional

autoscale (category)

Sets up a horizontal pod autoscaler to scale deployments automatically.

This configuration will automatically set up and associate the current deployment with a horizontal pod autoscaler in Kuberentes. Note that for this to work, you must also define resource limits and requests for a deployment otherwise the metrics-server will not be able to properly determine a deployments target CPU utilization.

autoscale.cpu_percent

The target CPU percent utilization before the horizontal pod autoscaler scales up a deployments replicas.

Type: int32
Optional

autoscale.max_replicas

The maximum amount of pods to scale to for a deployment.

Type: int32
Optional

autoscale.min_replicas

The minimum amount of pods to have for a deployment.

Type: int32
Optional

context

The kubectl context to use, as defined in the kubeconfig file.

Type: string
Optional

cpu (category)

Cpu resource configuration.

CPU lets you define resource limits and requests for a container in a deployment.

cpu.limit

Maximum amount of cpu to give the container. Supports m to indicate milli-cores.

Type: string
Optional

cpu.request

How much cpu to give the container in cpu cores. Supports m to indicate milli-cores.

Type: string
Optional

image_secret

Name of the Kubernetes secrete to use for the image.

This references an existing secret, waypoint does not create this secret.

Type: string
Optional

kubeconfig

Path to the kubeconfig file to use.

By default uses from current user's home directory.

Type: string
Optional
Environment Variable: KUBECONFIG

labels

A map of key value labels to apply to the deployment pod.

Type: map of string to string
Optional

memory (category)

Memory resource configuration.

Memory lets you define resource limits and requests for a container in a deployment.

memory.limit

Maximum amount of memory to give the container. Supports k for kilobytes, m for megabytes, and g for gigabytes.

Type: string
Optional

memory.request

How much memory to give the container in bytes. Supports k for kilobytes, m for megabytes, and g for gigabytes.

Type: string
Optional

namespace

Namespace to target deployment into.

Namespace is the name of the Kubernetes namespace to apply the deployment in. This is useful to create deployments in non-default namespaces without creating kubeconfig contexts for each.

Type: string
Optional

probe (category)

Configuration to control liveness and readiness probes.

Probe describes a health check to be performed against a container to determine whether it is alive or ready to receive traffic.

probe.failure_threshold

Number of times a liveness probe can fail before the container is killed.

FailureThreshold * TimeoutSeconds should be long enough to cover your worst case startup times.

Type: uint
Optional
Default: 30

probe.initial_delay

Time in seconds to wait before performing the initial liveness and readiness probes.

Type: uint
Optional
Default: 5

probe.timeout

Time in seconds before the probe fails.

Type: uint
Optional
Default: 5

probe_path

The HTTP path to request to test that the application is running.

Without this, the test will simply be that the application has bound to the port.

Type: string
Optional

replicas

The number of replicas to maintain.

If the replica count is maintained outside waypoint, for instance by a pod autoscaler, do not set this variable.

Type: int32
Optional

resources

A map of resource limits and requests to apply to a container on deploy.

Type: map of string to string
Optional

scratch_path

A path for the service to store temporary data.

A path to a directory that will be created for the service to store temporary data using EmptyDir.

Type: list of string
Optional

service_account

Service account name to be added to the application pod.

Service account is the name of the Kubernetes service account to add to the pod. This is useful to apply Kubernetes RBAC to the application.

Type: string
Optional

service_port

The TCP port that the application is listening on.

By default, this config variable is used for exposing a single port for the container in use. For multi-port configuration, use 'ports' instead.

Type: uint
Optional
Default: 3000

static_environment

Environment variables to control broad modes of the application.

Type: map of string to string
Optional

Output Attributes

Output attributes can be used in your waypoint.hcl as variables via artifact or deploy.

id

Type: string

name

Type: string

resource_state

Type: opaqueany.Any

kubernetes-apply (platform)

Deploy Kubernetes resources directly from a single file or a directory of YAML or JSON files.

This plugin lets you use any pre-existing set of Kubernetes resource files to deploy to Kubernetes. This plugin supports all the features of Waypoint. You may use Waypoint's templating features to template the resources with information such as the artifact from a previous build step, entrypoint environment variables, etc.

Requirements

This plugin requires "kubectl" to be installed since this plugin works by subprocessing to "kubectl apply". Other Waypoint Kubernetes plugins sometimes use the API directly but this plugin requires "kubectl".

"kubectl" must also be configured to access your Kubernetes cluster. You may specify an alternate kubeconfig file using the "kubeconfig" configuration parameter. If this isn't specified, the default kubectl lookup paths will be used.

Artifact Access

You may use Waypoint's templating features to access information such as the artifact from the build or push stages. An example below shows this by using templatedir mixed with variables such as artifact.image to dynamically configure the Docker image within a Kubernetes Deployment.

Entrypoint Functionality

Waypoint entrypoint functionality such as logs, exec, app configuration, and more require two properties to be true:

The running image must already have the Waypoint entrypoint installed and configured as the entrypoint. This should happen in the build stage.
Proper environment variables must be set so the entrypoint knows how to communicate to the Waypoint server. This step happens in this deployment stage.

Step 2 does not happen automatically. You must manually set the entrypoint environment variables using the templating feature. One of the examples below shows the entrypoint environment variables being injected.

URL Service

If you want your workload to be accessible by the Waypoint URL service, you must set the PORT environment variable within the pod with your web service and also be using the Waypoint entrypoint (documented in the previous section).

The PORT environment variable should be the port that your web service is listening on that the URL service will connect to. See one of the examples below for more details.

Interface

Examples

// The waypoint.hcl file
deploy {
  use "kubernetes-apply" {
    // Templated to perhaps bring in the artifact from a previous
    // build/registry, entrypoint env vars, etc.
    path        = templatedir("${path.app}/k8s")
    prune_label = "app=myapp"
  }
}

// ./k8s/deployment.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
  name: my-deployment
  labels:
    app: myapp
spec:
  replicas: 3
  selector:
    matchLabels:
      app: myapp
  template:
    metadata:
      labels:
        app: myapp
    spec:
      containers:
      - name: myapp
        image: ${artifact.image}:${artifact.tag}
        env:
          %{ for k,v in entrypoint.env ~}
          - name: ${k}
            value: "${v}"
          %{ endfor ~}

          # Ensure we set PORT for the URL service. This is only necessary
          # if we want the URL service to function.
          - name: PORT
            value: "3000"

Required Parameters

These parameters are used in the use stanza for this plugin.

path

Path to a file or directory of YAML or JSON files.

This will be used for kubectl apply to create a set of Kubernetes resources. Pair this with templatefile or templatedir templating functions to inject dynamic elements into your Kubernetes resources. Subdirectories are included recursively.

Type: string

prune_label

Label selector to prune resources that aren't present in the path.

This is a label selector that is used to search for any resources that are NOT present in the configured path and delete them.

Type: string

Optional Parameters

These parameters are used in the use stanza for this plugin.

context

The kubectl context to use, as defined in the kubeconfig file.

Type: string
Optional

kubeconfig

Path to the kubeconfig file to use.

If this isn't set, the default lookup used by kubectl will be used.

Type: string
Optional
Environment Variable: KUBECONFIG

Output Attributes

Output attributes can be used in your waypoint.hcl as variables via artifact or deploy.

prune_label

Type: string

kubernetes (releasemanager)

Manipulates the Kubernetes Service activate Deployments.

Interface

Required Parameters

These parameters are used in the use stanza for this plugin.

ingress (category)

Configuration to set up an ingress resource to route traffic to the given application from an ingress controller.

An ingress resource can be created on release that will route traffic to the Kubernetes service. Note that before this happens, the Kubernetes cluster must already be configured with an Ingress controller. Otherwise there won't be a way for inbound traffic to be routed to the ingress resource.

ingress.annotations

Annotations to be applied to the ingress resource.

Type: map of string to string
Optional

ingress.default

Sets the ingress resource to be the default backend for any traffic that doesn't match existing ingress rule paths.

Type: bool
Optional
Default: false

ingress.host

If set, will configure the ingress resource to have the ingress controller route traffic for any inbound requests that match this host. IP addresses are not allowed, nor are ':' delimiters. Wildcards are allowed to a certain extent. For more details check out the Kubernetes documentation.

Type: string
Optional

ingress.path

The route rule that should be used to route requests to this ingress resource. A path must begin with a '/'.

Type: string
Optional
Default: /

ingress.path_type

Defines the kind of rule the path will be for the ingress controller. Valid path types are 'Exact', 'Prefix', and 'ImplementationSpecific'.

Type: string
Optional
Default: Prefix

ingress.tls (category)

A stanza of TLS configuration options for traffic to the ingress resource.

ingress.tls.hosts

A list of hosts included in the TLS certificate.

ingress.tls.secret_name

The Kubernetes secret name that should be used to look up or store TLS configs.

Optional Parameters

These parameters are used in the use stanza for this plugin.

annotations

Annotations to be applied to the kube service.

Type: map of string to string
Optional

context

The kubectl context to use, as defined in the kubeconfig file.

Type: string
Optional

kubeconfig

Path to the kubeconfig file to use.

By default uses from current user's home directory.

Type: string
Optional
Environment Variable: KUBECONFIG

load_balancer

Indicates if the Kubernetes Service should LoadBalancer type.

If the Kubernetes Service is not a LoadBalancer and node_port is not set, then the Service uses ClusterIP.

Type: bool
Optional

namespace

Namespace to create Service in.

Namespace is the name of the Kubernetes namespace to create the deployment in This is useful to create Services in non-default namespaces without creating kubeconfig contexts for each.

Type: string
Optional

node_port

The TCP port that the Service should consume as a NodePort.

If this is set but load_balancer is not, the service will be NodePort type, but if load_balancer is also set, it will be LoadBalancer.

Type: uint
Optional

port

The TCP port that the application is listening on.

Type: uint
Optional
Default: 80

ports

A map of ports and options that the application is listening on.

Used to define and configure multiple ports that the application is listening on. Available keys are 'port', 'node_port', 'name', and 'target_port'. If 'node_port' is set but 'load_balancer' is not, the service will be NodePort type. If 'load_balancer' is also set, it will be LoadBalancer. Ports defined will be TCP protocol. Note that 'name' is required if defining more than one port.

Type: list of map of string to string
Optional

kubernetes (configsourcer)

Read configuration values from Kubernetes ConfigMap or Secret resources. Note that to read a config value from a Secret, you must set secret = true. Otherwise Waypoint will load a dynamic value from a ConfigMap.

Examples

config {
  env = {
    PORT = dynamic("kubernetes", {
      name = "my-config-map"
      key = "port"
    })

    DATABASE_PASSWORD = dynamic("kubernetes", {
      name = "database-creds"
      key = "password"
      secret = true
    })
  }
}

Required Parameters

These parameters are used in dynamic for sourcing configuration values or input variable values.

key

The key in the ConfigMap or Secret to read the value from.

ConfigMaps and Secrets store data in key/value format. This specifies the key to read from the resource. If you want multiple values you must specify multiple dynamic values.

Type: string

name

The name of the ConfigMap of Secret.

Type: string

Optional Parameters

These parameters are used in dynamic for sourcing configuration values or input variable values.

namespace

The namespace to load the ConfigMap or Secret from.

By default this will use the namespace of the running pod. If this config source is used outside of a pod, this will use the namespace from the kubeconfig.

Type: string
Optional

secret

This must be set to true to read from a Secret. If it is false we read from a ConfigMap.

Type: bool
Optional

kubernetes (task)

Launch a Kubernetes pod for on-demand tasks from the Waypoint server.

This will use the standard Kubernetes environment variables to source authentication information for Kubernetes. If this is running within Kubernetes itself (typical for a Kubernetes-based installation), it will use the pod's service account unless other auth is explicitly given. This allows the task launcher to work by default.

Interface

Examples

task {
    use "kubernetes" {}
}

Required Parameters

These parameters are used in the use stanza for this plugin.

cpu

Cpu resource request to be added to the task container.

Type: k8s.ResourceConfig

ephemeral_storage

Ephemeral_storage resource request to be added to the task container.

Type: k8s.ResourceConfig

memory

Memory resource request to be added to the task container.

Type: k8s.ResourceConfig

Optional Parameters

These parameters are used in the use stanza for this plugin.

context

The kubectl context to use, as defined in the kubeconfig file.

Type: string
Optional

image_pull_policy

Pull policy to use for the task container image.

Type: string
Optional

image_secret

Name of the Kubernetes secret to use for the image.

This references an existing secret; Waypoint does not create this secret.

Type: string
Optional

kubeconfig

Path to the kubeconfig file to use.

By default uses from current user's home directory.

Type: string
Optional
Environment Variable: KUBECONFIG

namespace

Namespace in which to launch task.

Type: string
Optional

service_account

Service account name to be added to the application pod.

Service account is the name of the Kubernetes service account to add to the pod. This is useful to apply Kubernetes RBAC to the application.

Type: string
Optional

watchtask_startup_timeout_seconds

This option configures how long the WatchTask should wait for a task pod to start-up before attempting to stream its logs. If the pod does not start up within the given timeout, WatchTask will exit.

Type: int
Optional
Default: 30

Output Attributes

Output attributes can be used in your waypoint.hcl as variables via artifact or deploy.

id

Type: string