benthos-helm-chart

This chart is relatively immature, and may not yet support all workloads or benthos features. Please feel free to test and provide feedback, report bugs, and/or contribute!

Repository

To add this repo:

helm repo add benthos https://benthosdev.github.io/benthos-helm-chart/

Then helm search repo benthos for all charts.

Versions

Benthos is currently on v4. If you need v3 for any reason, please set image.tag in your values.yaml.

Configuration

Benthos Parameters

For more information on configuring the HTTP component, refer to the Benthos HTTP component documentation.

Name	Description	Value
image.repository	Docker image repository	benthosdev/benthos
image.pullPolicy	Docker image pull policy	IfNotPresent
image.tag	Docker image tag override	""
imagePullSecrets	Docker registry secrets array	[]
nameOverride	Sets name override	""
fullnameOverride	Sets full name override	""
serviceaccount.create	Enables creation of serviceaccount	false
serviceaccount.annotations	Sets serviceaccount annotations	{}
serviceaccount.name	Sets serviceaccount name	""
podAnnotations	Sets pod annotations	{}
podLabels	Sets pod labels	{}
commonLabels	Add labels that are common for all k8s components	""
podSecurityContext	Sets pod security context	{}
command	Replaces entrypoint command of benthos docker	""
args	Override default arguments passed with `command`	""
securityContext	Sets security context	{}
service.type	Kubernetes service type	ClusterIP
service.ports	Kubernetes service ports	[]
ingress.enabled	Enables ingress	false
ingress.className	Sets ingress class name	""
ingress.annotations	Sets ingress annotations	{}
ingress.tls	Sets ingress TLS configuration	[]
ingress.hosts	Sets ingress hosts configuration	[]
env	Sets benthos environment variables	[]
updateStrategy	Add Deployment Strategy	{}
resources	Set pod resource limits and/or requests	{}
autoscaling.enabled	Enables the horizontal pod autoscaler	false
autoscaling.minReplicas	Sets min number of replicas	1
autoscaling.maxReplicas	Sets max numbers of replicas	100
autoscaling.targetCPUUtilizationPercentage	Sets desired CPU autoscaling threshold	80
autoscaling.targetMemoryUtilizationPercentage	Sets desired memory autoscaling threshold	80
nodeSelector	Sets node selector configuration	{}
tolerations	Sets tolerations configuration	[]
affinity	Sets affinity configuration	{}
extraVolumes	Defines extra volumes configuration	[]
extraVolumeMounts	Sets additional mounts defined in extraVolumes	[]
streams.enabled	Enables 'Benthos streams' mode	false
streams.streamsConfigMap	Name of K8s configMap containing streams configs	""
streams.api.enable	Enables streams API	true
http.enabled	Enables the HTTP server component	true
http.address	HTTP server component binding address	0.0.0.0:4195
http.rootPath	General Benthos HTTP endpoint prefix	/benthos
http.debugEndpoints	Enables debugging endpoints	false
http.cors.enabled	Enables Cross-Origin Resource Sharing	false
http.cors.allowedOrigins	Allowed source domains for CORS	""
http.tls.enabled	Enables TLS for all Benthos endpoints	false
http.tls.secretName	`kubernetes.io/tls` secret name	""
podDisruptionBudget.enabled	Enables Pod Disruption Budget for Benthos	false
podDisruptionBudget.minAvailable	Min number of pods available at all times	""
podDisruptionBudget.maxUnavailable	Max unavailable number of pods	""
watch	Enables watch mode	false
initContainers	Add any custom init container	[]
config	Benthos component configuration	""

Config

The config parameter should contain the configuration as it would be parsed by the Benthos binary.

For example, the default Helm chart config block looks like this:

# /benthos.yaml configuration
config: |-
  input:
    label: "no_config_in"
    generate:
      mapping: root = "This Benthos instance is unconfigured!"
      interval: 1m
  output:
    label: "no_config_out"
    stdout:
      codec: lines

Adding an http block here is not recommended, please use the Helm directives described above.

TLS

Benthos can be instructed to serve all endpoints exlusively over HTTPS. This means that TLS configured in this way is not terminated at an ingress controller, but handled "end-to-end" at the container/binary. Prerequisites to enable TLS:

Set service.port to 443 in values.yaml
Create a Kubernetes secret in the targeted namespace of type kubernetes.io/tls

When TLS is enabled, the Kubernetes readiness and liveness probes will operate over HTTPS to the same container port (default 4195).

Streams mode

When running Benthos in streams mode, all individual stream configuration files should be combined and placed in a single Kubernetes configMap, like so:

apiVersion: v1
kind: ConfigMap
metadata:
  name: benthos-streams
data:
  hello.yaml: |
    input:
      generate:
        mapping: root = "woof"
        interval: 5s
        count: 0
    output:
      stdout:
        codec: lines
  aaaaa.yaml: |
    input:
      generate:
        mapping: root = "meow"
        interval: 2s
        count: 0
    output:
      stdout:
        codec: lines

Setting streams.enabled: true in values.yaml adds streams /streams/*.yaml to the pod command arguments.

This exposes a streams API by default, where configuration can be viewed and altered by anyone with access to the configured benthos endpoint. As of benthos version 3.61.0, the streams API can be disabled to prevent changes to the configuration[s] of a running Benthos in streams mode via config files; use .Values.streams.api.enabled = false to toggle the API off (this adds an additional argument to the pod command to become streams --no-api /streams/*.yaml).

With streams mode enabled, /benthos.yaml should only contain the configuration for http as defined either by the http values block or the defaults; the contents of the streams configMap (as defined and pulled into the deployment by the name set in streams.streamsConfigMap) are deployed to /streams/*.yaml. Additionally, general config like metrics, logger , and tracing.

From the benthos streams --help context:

In streams mode the stream fields of a root target config (input, buffer,
      pipeline, output) will be ignored. Other fields will be shared across all
      loaded streams (resources, metrics, etc).

Currently this streams configMap should be applied separately from and before installation of the helm chart; support for deploying additional configMaps within the chart may be implemented later.

and1truong / benthos-helm-chart