This guide explains how to install Grafana Mimir on a Kubernetes cluster using Helm.
Installation Steps:
1.Install Mimir-distributed via Helm
2.Install kube-prometheus-stack via Helm
3.Verify that Prometheus instances are scraping samples from various targets and pushing them to Grafana Mimir using Prometheus' remote write API.
4.Configure a ServiceMonitor for Prometheus, labeling release: kube-prometheus-stack, to allow the Prometheus operator to discover which exporter services to scrape.
5.Verify that Mimir metrics are being scraped through the Prometheus interface.
6.Follow the instructions in the Grafana Mimir documentation to build the Mimir dashboards.
1.Install Mimir-distributed via Helm
Add Mimir Helm repository:
helm repo add mimir-test grafana/mimir-distributed
Install Mimir using Helm: Create a values file (let's name it mimir.yaml) with the mimir configuration. Here's an example of how it might look:
metaMonitoring:
serviceMonitor:
enabled: true
labels:
release: kube-prometheus-stack
mimir:
structuredConfig:
alertmanager_storage:
s3:
access_key_id: access_key
bucket_name: mimir-alert
endpoint: s3.eu-central-1.amazonaws.com
secret_access_key: secret_access
blocks_storage:
backend: s3
bucket_store:
sync_dir: /data/tsdb-sync
s3:
access_key_id: access_key
bucket_name: mimir-tsdb
endpoint: s3.eu-central-1.amazonaws.com
secret_access_key: secret_access
tsdb:
dir: /data/tsdb
compactor:
data_dir: /data
frontend:
align_queries_with_step: true
log_queries_longer_than: 10s
ingester:
instance_limits:
max_ingestion_rate: 0
ring:
final_sleep: 0s
num_tokens: 512
ingester_client:
grpc_client_config:
max_recv_msg_size: 104857600
max_send_msg_size: 104857600
limits:
ingestion_rate: 40000
max_global_series_per_metric: 0
max_global_series_per_user: 0
memberlist:
abort_if_cluster_join_fails: false
compression_enabled: false
ruler:
alertmanager_url: dnssrvnoa+http://_http-metrics._tcp.{{ template "mimir.fullname"
. }}-alertmanager-headless.{{ .Release.Namespace }}.svc.{{ .Values.global.clusterDomain
}}/alertmanager
enable_api: true
rule_path: /data
ruler_storage:
s3:
access_key_id: access_key
bucket_name: mimir-ruler-test
endpoint: s3.eu-central-1.amazonaws.com
secret_access_key: secret_access
runtime_config:
file: /var/{{ include "mimir.name" . }}/runtime.yaml
server:
grpc_server_max_concurrent_streams: 1000
grpc_server_max_recv_msg_size: 104857600
grpc_server_max_send_msg_size: 104857600
minio:
enabled: false
querier:
replicas: 3
Update Helm repository within the mimir-test namespace:
helm repo update -n mimir-test
install Mimir using Helm, specifying custom values from mimir.yaml:
helm -n mimir-test upgrade mimir-test grafana/mimir-distributed --values=mimir.yaml --install
This setup will deploy mimir with S3 storage backend and kube-prometheus-stack configured to send metrics to mimir. Adjust configurations as per your specific environment and requirements.
Object storage: I configured three separate buckets named mimir-alert, mimir-ruler-test, and mimir-tsdb
After installing the Helm chart and waiting awhile, I could see data starting to show up in the bucket from my object store web interface:
2.Install kube-prometheus-stack via Helm
Install kube-prometheus-stack: Create a values file (let's name it kube-prometheus-stack-values.yaml) for kube-prometheus-stack:
The role of Prometheus Prometheus instances scrape samples from various targets and push them to Grafana Mimir by using Prometheus’ remote write API. ServiceMonitor for Prometheus
label release: kube-prometheus-stack, which is what allows the Prometheus operator to discover which exporter services to scrape. I verified through the Prometheus interface that the Mimir metrics were being scraped. Then I followed the instructions in the Grafana Mimir documentation to build the Mimir dashboards.
vi kube-prometheus-stack-values.yaml
prometheus:
prometheusSpec:
remoteWrite:
- url: http://mimir-test-nginx.mimir-test.svc:80/api/v1/push
externalLabels:
environment: mimir
install kube-prometheus-stack using Helm:
helm install kube-prometheus-stack prometheus-community/kube-prometheus-stack -f kube-prometheus-stack-values.yaml --namespace monitoring --create-namespace
Access Grafana Dashboard
Port Forward Grafana Service
kubectl port-forward svc/kube-prometheus-stack-grafana 3000:80 -n monitoring
Access Grafana Dashboard Open a web browser and go to localhost:3000.
Login with the following credentials: Username: admin Password: (retrieve password using the following command)
kubectl get secret prom-grafana -o jsonpath="{.data.admin-password}" -n monitoring | base64 --decode ; echo
Username: admin Password: prom-operator
Add mimir datasource in grafana with remote read URL http://mimir-test-nginx.mimir-test.svc:80/prometheus
Add default mimir (mimir mixin) dashboards to grafana
