gNMIc Cluster

This project enables users to run an autoscaling gNMIc cluster on kubernetes. gNMIc is an openconfig project that allows users to stream telemetry from gNMI enabled routers.

This project was co-funded by GEANT in the GN5-1 WP6 Incubator.

Telemetry pipeline architecture

Prerequisites

• Kubernetes cluster

• A strimzi operator and kafka cluster deployment within your k8s cluster Strimzi

• Minimum CPU and RAM for 300 Routers and subscription to all metrics for processing +/- 2 billion events per day

  CPU	Memory	Storage (90 days retention)
  20	160Gi	3Ti

Installation

helm repo add gnmic-cluster-chart https://workfloworchestrator.org/gnmic-cluster-chart/
helm repo update
helm upgrade --install gnmic-cluster gnmic-cluster-chart/gnmic-cluster

Helm Chart Documentation

gnmic-cluster

This chart deploys a gNMIc cluster. This application enables users to stream telemetry from gNMI capable routers.

Homepage: https://workfloworchestrator/gnmic-cluster-chart

Maintainers

Name	Email	Url
Workflow Orchestrator Programme	[email protected]	https://workfloworchestrator.org

Source Code

• https://github.com/workfloworchestrator/gnmic-cluster-chart/tree/main/charts/gnmic

Requirements

Repository	Name	Version
https://emberstack.github.io/helm-charts	reflector	7.1.288
oci://registry-1.docker.io/bitnamicharts	redis	20.1.5

Values

Key	Type	Default	Description
affinity	object	{}	Affinity for the k8s scheduler, this is a mapping per collector and relay
autoscaling	object	{"enabled":true,"locker":"consul","maxReplicas":30,"minReplicas":4,"targetCPUUtilizationPercentage":80}	Autscaling configuration, enabled by default to maximise the efficiency of collection.
autoscaling.enabled	bool	true	Autoscaling enabled
autoscaling.locker	string	"consul"	'consul' or 'k8s'
autoscaling.maxReplicas	int	30	Autoscaling upper bound
autoscaling.minReplicas	int	4	Autoscaling lower bound
autoscaling.targetCPUUtilizationPercentage	int	80	targetCPUUtilizationPercentage or targetMemoryUtilizationPercentage
config	object	{"collector":{},"global":{"api-server":{"address":":7890","debug":false,"enable-metrics":true},"debug":false,"encoding":"proto","format":"event","gzip":false,"password":"${GNMIC_PASSWORD}","skip-verify":true,"username":"${GNMIC_USERNAME}"},"relay":{}}	gNMIc configuration
config.collector	object	{}	Collector configuration, must be provided in yaml by following the instructions here: https://gnmic.openconfig.net/user_guide/subscriptions/
config.global	object	{"api-server":{"address":":7890","debug":false,"enable-metrics":true},"debug":false,"encoding":"proto","format":"event","gzip":false,"password":"${GNMIC_PASSWORD}","skip-verify":true,"username":"${GNMIC_USERNAME}"}	Global gNMIc Configuration extra keys can be added or removed according to: https://gnmic.openconfig.net/user_guide/configuration_file/#global-flags
config.global.password	string	"${GNMIC_PASSWORD}"	An example of using environment value variables
config.global.username	string	"${GNMIC_USERNAME}"	An example of using environment value variables
config.relay	object	{}	Relay configuration, must be provided in yaml by following the instructions here: https://gnmic.openconfig.net/user_guide/outputs/
fullnameOverride	string	""	Fullname override
image.pullPolicy	string	"Always"	Image pull policy.
image.repository	string	"ghcr.io/openconfig/gnmic"	Image repository.
image.tag	string	"latest"	Overrides the image tag whose default is the chart appVersion.
imagePullSecrets	list	[]	Required for private image repository
ingress.annotations	object	{}	Ingress annotations
ingress.className	string	""	The ingress class name
ingress.enabled	bool	false	The ingress for the Api
ingress.hosts	list	[]	The ingress host rules
ingress.tls	list	[]	TLS rules
kafka	object	{"bootstrap_server":"","cluster":"","enabled":true,"group":"gnmic-collector","kafka_namespace":"","topic":{"config":{},"name":"gnmic-telemetry","partitions":15,"replicas":3},"user":{"mechanism":"scram-sha-512","name":"gnmic"}}	Requires a Strimzi Operator inside the cluster and an active Kafka deployment https://strimzi.io/documentation/
kafka.bootstrap_server	string	""	Kafka Bootstrap address: fqdn + port. eg: kafka-bootstrap.production.svc.cluster.local:9094
kafka.cluster	string	""	The name of the cluster inside the
kafka.enabled	bool	true	Enabling this will create the strimzi artefacts to create the kafka objects in K8s. This will also configure the secret reflector so the kafka user secret can be
kafka.group	string	"gnmic-collector"	The kafka user group
kafka.kafka_namespace	string	""	The namespace where the kafka manifests need to be installed so the operator can create the user topics and rolebindings
kafka.topic	object	{"config":{},"name":"gnmic-telemetry","partitions":15,"replicas":3}	Kafka topic configuration
kafka.topic.config	object	{}	Kafka Topic configuration
kafka.topic.name	string	"gnmic-telemetry"	Topic name
kafka.topic.partitions	int	15	Minumum number of partitions (cruise control may scale to more)
kafka.topic.replicas	int	3	Minimum number of replicas ( num(replicas) =< num(kafka_nodes)
kafka.user	object	{"mechanism":"scram-sha-512","name":"gnmic"}	The kafka user configuration
nameOverride	string	""	Name override
podAnnotations	object	{}	Any extra pod annotations
podLabels	object	{}	Any extra pod annotations
podSecurityContext	object	{}	The Custom security context written in Yaml
redis	object	{"enabled":false}	Optional deployment of redis inside the namespace. This
redis.enabled	bool	false	Enable or disable usage of the Bitnami redis
reflector	object	{"enabled":false}	Optional deployment of the secret reflector. This is necesarry when running the
replicaCount	int	3	Number of replicas for the Collector and Relay statefulset. This will be overridden when autoscaling is activated
resources.collector	object	{}	Resources for the collector, mandatory for functional autoscaling
resources.consul	object	{}	Resources for the consul deployment
resources.relay	object	{}	Resources for the relay, mandatory for functional autoscaling
secrets	object	{"name":"gnmic-secret","opaque":{"enabled":false,"items":{}},"secret_provider":{"config":{"keyvault":"","nodePublishSecretRefName":"","provider":"","tenantId":""},"enabled":false,"objects":[]}}	Secret configuration
secrets.name	string	"gnmic-secret"	Name of the secret
secrets.opaque	object	{"enabled":false,"items":{}}	Usimg an opaque secret inside kubernets
secrets.opaque.enabled	bool	false	Whether it is enabled
secrets.opaque.items	object	{}	Secret items
secrets.secret_provider	object	{"config":{"keyvault":"","nodePublishSecretRefName":"","provider":"","tenantId":""},"enabled":false,"objects":[]}	Using a secret provider allows the user to inject secrets from a vault or cloud provider
secrets.secret_provider.config	object	{"keyvault":"","nodePublishSecretRefName":"","provider":"","tenantId":""}	K8s secret provider configuration
secrets.secret_provider.config.keyvault	string	""	Name of the keyvault
secrets.secret_provider.config.nodePublishSecretRefName	string	""	Secret with credentials to access the keyvault
secrets.secret_provider.config.provider	string	""	eg: azure
secrets.secret_provider.config.tenantId	string	""	Tenant id
secrets.secret_provider.enabled	bool	false	Secret provider enabled
secrets.secret_provider.objects	list	[]	List of objects to retrieve from the keyvault. These objects will be mapped to enviroment variables inside the containers
server	object	{"cache":{"address":"","type":"redis"},"enabled":true,"port":57400}	gNMI server capabilities: https://gnmic.openconfig.net/user_guide/gnmi_server/
server.cache	object	{"address":"","type":"redis"}	Override of the default cache
server.cache.address	string	""	Redis cache address in URI form, e.g: redis-master.production.svc.cluster.local:6379
server.cache.type	string	"redis"	Only redis is supported
server.enabled	bool	true	Whether the server is enabled
server.port	int	57400	Default port
service.annotations	object	{}	The extra service annotations to be compatible with clouds
service.externalTrafficPolicy	string	"Cluster"	The ExternalTrafficPolicy
service.port	int	7890	The port being exposed
service.type	string	"ClusterIP"	The type of service
serviceAccount.annotations	object	{}	Annotations to add to the service account
serviceAccount.automount	bool	true	Automatically mount a ServiceAccount's API credentials?
serviceAccount.create	bool	true	Specifies whether a service account should be created, must be set to true when using the k8s locker
serviceAccount.name	string	""	The name of the service account to use. If not set and create is true, a name is generated using the fullname template
servicemonitor.enabled	bool	false	For monitoring by Prometheus
tolerations	object	{}	Tolerations for the k8s scheduler, this is a mapping per collector and relay

---

Autogenerated from chart metadata using helm-docs v1.14.2

Updating documentation

The Helm chart and readme are documented separately.

Helm

The Helm chart is documented use helm-docs Link Run it as follows:

helm-docs .

Readme generation

The README generation is done by editing the README.md.tpl file. and running the command

md_incl README.md.tpl README.md

This assumes you have md_incl installed. link