On this page:

Log Analytics (Elasticsearch and Kibana)

In June 2020, Elastic announced that starting from 7.8 release they will provide multi-architecture docker images supporting AMD64 and ARM64 architectures.

To facilitate the deployment on a Kubernetes cluster ECK project was created. ECK (Elastic Cloud on Kubernetes) automates the deployment, provisioning, management, and orchestration of ELK Stack (Elasticsearch, Kibana, APM Server, Enterprise Search, Beats, Elastic Agent, and Elastic Maps Server) on Kubernetes based on the operator pattern.

ECK Operator will be used to deploy Elasticsearh and Kibana.

ECK Operator installation

Step 1: Add the Elastic repository:

helm repo add elastic https://helm.elastic.co

Step2: Fetch the latest charts from the repository:
```
helm repo update
```
Step 3: Create namespace
```
kubectl create namespace logging
```

Step 3: Install ECK operator in the logging namespace

helm install elastic-operator elastic/eck-operator --namespace logging

Step 4: Monitor operator logs:

kubectl -n logging logs -f statefulset.apps/elastic-operator

Elasticsearch installation

Basic instructions can be found in ECK Documentation: “Deploy and elasticsearch cluster”

Step 1: Create a manifest file containing basic configuration: one node elasticsearch using Longhorn as storageClass and 5GB of storage in the volume claims.
```
apiVersion: elasticsearch.k8s.elastic.co/v1
kind: Elasticsearch
metadata:
  name: efk
  namespace: logging
spec:
  version: 8.1.2
  nodeSets:
  - name: default
    count: 1    # One node elastic search cluster
    config:
      node.store.allow_mmap: false # Disable memory mapping
    volumeClaimTemplates: 
      - metadata:
          name: elasticsearch-data
        spec:
          accessModes:
          - ReadWriteOnce
          resources:
            requests:
              storage: 5Gi
          storageClassName: longhorn
  http:
    tls: # Disabling TLS automatic configuration. Note(3)
      selfSignedCertificate:
        disabled: true
```
- About Virtual Memory configuration (mmap)
  
  By default, Elasticsearch uses memory mapping (mmap) to efficiently access indices. To disable this default mechanism add the following configuration option:
```
node.store.allow_nmap: false
```
  Usually, default values for virtual address space on Linux distributions are too low for Elasticsearch to work properly, which may result in out-of-memory exceptions. This is why mmap is disable.
  
  For production workloads, it is strongly recommended to increase the kernel setting vm.max_map_count to 262144 and leave node.store.allow_mmap unset.
  
  See further details in ECK Documentation: “Elastisearch Virtual Memory”
- About Persistent Storage configuration
  
  Longhorn is configured for Elastisearch POD’s persistent volumes
```
volumeClaimTemplates:
  - metadata:
      name: elasticsearch-data
    spec:
      accessModes:
      - ReadWriteOnce
      resources:
        requests:
          storage: 5Gi
      storageClassName: longhorn
```
  See how to configure PersistenVolumeTemplates for Elasticsearh using this operator in ECK Documentation: “Volume claim templates”
- Disable TLS configuration
```
http:
  tls:
    selfSignedCertificate:
      disabled: true
```
  By default ECK configures secured communications with auto-signed SSL certificates. Access to its service endpoint on port 9200 is only available through https.
  
  Disabling TLS automatic configuration in Elasticsearch HTTP server enables Cluster Service Mesh to gather more statistics about connections. Service Mesh is parsing plain text traffic (HTTP) instead of encrypted (HTTPS).
  
  Cluster service mesh will enforce secure communications using TLS between all PODs.
- About limiting resources assigned to ES
  
  In Kubernetes, limits in the consumption of resources (CPU and memory) can be assigned to PODs. See “Kubernetes Doc - Resource Management for Pods and Containers”.
  
  resource requests defines the minimum amount of resources that must be available for a Pod to be scheduled; resource limits defines the maximum amount of resources that a Pod is allowed to consume.
  
  When you specify the resource request for containers in a Pod, the kube-scheduler uses this information to decide which node to place the Pod on. When you specify a resource limit for a container, the kubelet enforces those limits so that the running container is not allowed to use more of that resource than the limit you set. The kubelet also reserves at least the request amount of that system resource specifically for that container to use.
  
  In case of using ECK Operator is it recommended to specify those resource limits and resource request to each of the Objects created by the Operator. See details on how to setup those limits in ECK Documentation - Manage compute resources.
  
  For example memory heap assigned to JVM is calculated based on that resource limits, “The heap size of the JVM is automatically calculated based on the node roles and the available memory. The available memory is defined by the value of resources.limits.memory set on the elasticsearch container in the Pod template, or the available memory on the Kubernetes node is no limit is set”.
  
  By default, ECK does not specify any limit to CPU resource and it defines resources.limits.memory for ElasticSearch POD set to 2GB.
  
  In production environment this default limit should be increased. In lab environments where memory resources are limited it can be decreased to reduce ES memory footprint.
  
  In both scenarios, the limit can be changed in in Elasticsearch object (podTemplate section).
```
  podTemplate:
    # Limiting Resources consumption
    spec:
      containers:
      - name: elasticsearch
        resources:
          requests:
            memory: 1Gi
          limits:
            memory: 1Gi
```
Step 2: Apply manifest
```
kubectl apply -f manifest.yml
```
Step 3: Check Elasticsearch status
```
kubectl get elasticsearch -n logging
NAME   HEALTH   NODES   VERSION   PHASE   AGE
efk    yellow   1       8.1.2    Ready   139m
```
Note:

Elasticsearch status HEALTH=yellow indicates that only one node of the Elasticsearch is running (no HA mechanism), PHASE=Ready indicates that the server is up and running

Elasticsearch authentication

By default ECK configures user authentication to access elasticsearch service. ECK defines a default admin esaticsearch user (elastic) and with a password which is stored within a kubernetes Secret.

Both to access elasticsearch from Kibana GUI or to configure Fluentd collector to insert data, elastic user/password need to be provided.

Password is stored in a kubernetes secret (<efk_cluster_name>-es-elastic-user). Execute this command for getting the password

kubectl get secret -n logging efk-es-elastic-user -o=jsonpath='{.data.elastic}' | base64 --decode; echo

Recently ECK has added support to define additional custom users and roles. Custom users are added ES File-based Authentication.

See Users and roles from elastic cloud-on-k8s documentation.

To allow fluentd and prometheus exporter to access our elasticsearch cluster, we can define two role that grants the necessary permission for the two users we will be creating (fluentd, prometheus).

Step 1: Create Secrets containing roles definitions

Fluentd user role:

kind: Secret
apiVersion: v1
metadata:
  name: es-fluentd-roles-secret
  namespace: logging
stringData:
  roles.yml: |-
    fluentd_role:
      cluster: ['manage_index_templates', 'monitor', 'manage_ilm']
      indices:
      - names: [ '*' ]
        privileges: [
          'indices:admin/create',
          'write',
          'create',
          'delete',
          'create_index',
          'manage',
          'manage_ilm'
        ]

Prometheus Exporter user role:

kind: Secret
apiVersion: v1
metadata:
  name: es-prometheus-roles-secret
  namespace: logging
stringData:
  roles.yml: |-
    prometheus_role:
      cluster: [
        'cluster:monitor/health',
        'cluster:monitor/nodes/stats',
        'cluster:monitor/state',
        'cluster:monitor/nodes/info',
        'cluster:monitor/prometheus/metrics'
      ] 
      indices:
      - names: [ '*' ]
        privileges: [ 'indices:admin/aliases/get', 'indices:admin/mappings/get', 'indices:monitor/stats', 'indices:data/read/search' ]

Step 2. Create the Secrets containing user name, password and mapped role

Fluentd user:

apiVersion: v1
kind: Secret
metadata:
  name: es-fluentd-user-file-realm
  namespace: logging
type: kubernetes.io/basic-auth
data:
  username: <`echo -n 'fluentd' | base64`>
  password: <`echo -n 'supersecret' | base64`>
  roles: <`echo -n 'fluentd_role' | base64`>

Prometheus exporter user:

apiVersion: v1
kind: Secret
metadata:
  name: es-prometheus-user-file-realm
  namespace: logging
type: kubernetes.io/basic-auth
data:
  username: <`echo -n 'prometheus' | base64`>
  password: <`echo -n 'supersecret' | base64`>
  roles: <`echo -n 'prometheus_role' | base64`>

Step 3: Modify Elasticsearch yaml file created in step 1 of ES installation.

Add the following lines to ElasticSearch manifest file:

apiVersion: elasticsearch.k8s.elastic.co/v1
kind: Elasticsearch
metadata:
  name: efk
  namespace: logging
spec:
  auth:
    roles:
    - secretName: es-fluentd-roles-secret
    - secretName: es-prometheus-roles-secret
    fileRealm:
    - secretName: es-fluentd-user-file-realm
    - secretName: es-prometheus-user-file-realm
...

In addition to the elastic user we can also create an super user account for us to login, we can create the account just like how we created the fluentd or prometheus user, but instead with the role set to superuser.

Accesing Elasticsearch from outside the cluster

By default Elasticsearh HTTP service is accesible through Kubernetes ClusterIP service types (only available within the cluster). To make it available outside the cluster Ingress NGINX can be configured to enable external communication with Elasicsearh server.

This exposure will be useful for doing remote configurations on Elasticsearch through its API from pimaster node. For example: to configure backup snapshots.

Step 1. Create the ingress rule manifest

---
# HTTPS Ingress
apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
  name: elasticsearch-ingress
  namespace: logging
  annotations:
    # Enable cert-manager to create automatically the SSL certificate and store in Secret
    cert-manager.io/cluster-issuer: ca-issuer
    cert-manager.io/common-name: elasticsearch.picluster.ricsanfre.com
spec:
  ingressClassName: nginx
  tls:
    - hosts:
        - elasticsearch.picluster.ricsanfre.com
      secretName: elasticsearch-tls
  rules:
    - host: elasticsearch.picluster.ricsanfre.com
      http:
        paths:
          - path: /
            pathType: Prefix
            backend:
              service:
                name: efk-es-http
                port:
                  number: 9200

ingress NGINX exposes elasticsearch server as elasticsearch.picluster.ricsanfre.com virtual host, routing rules are configured for redirecting all incoming HTTP traffic to HTTPS and TLS is enabled using a certificate generated by Cert-manager.

See Ingress NGINX configuration document for furher details.

Step 2: Apply manifest
```
kubectl apply -f manifest.yml
```

Step 3. Access to Elastic HTTP service

UI can be access through http://elasticsearch.picluster.ricsanfre.com using loging elastic and the password stored in <efk_cluster_name>-es-elastic-user.

It should shows the following output (json message)

{
  "name" : "efk-es-default-0",
  "cluster_name" : "efk",
  "cluster_uuid" : "w5BUxIY4SKOtxPUDQfb4lQ",
  "version" : {
    "number" : "8.1.2",
    "build_flavor" : "default",
    "build_type" : "docker",
    "build_hash" : "31df9689e80bad366ac20176aa7f2371ea5eb4c1",
    "build_date" : "2022-03-29T21:18:59.991429448Z",
    "build_snapshot" : false,
    "lucene_version" : "9.0.0",
    "minimum_wire_compatibility_version" : "7.17.0",
    "minimum_index_compatibility_version" : "7.0.0"
  },
  "tagline" : "You Know, for Search"
}

Kibana installation

Step 1. Create a manifest file

apiVersion: kibana.k8s.elastic.co/v1
kind: Kibana
metadata:
  name: kibana
  namespace: logging
spec:
  version: 8.1.2
  count: 2 # Elastic Search statefulset deployment with two replicas
  elasticsearchRef:
    name: "efk"
  http:  # NOTE disabling kibana automatic TLS configuration
    tls:
      selfSignedCertificate:
        disabled: true

Step 2: Apply manifest
```
kubectl apply -f manifest.yml
```

Step 3: Check kibana status

kubectl get kibana -n logging
NAME   HEALTH   NODES   VERSION   AGE
efk    green    1       8.1.2    171m

Note:

Kibana status HEALTH=green indicates that Kibana is up and running.

Ingress rule for Kibana

Make accesible Kibana UI from outside the cluster through Ingress Controller

Step 1. Create the ingress rule manifest

---
# HTTPS Ingress
apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
  name: kibana-ingress
  namespace: logging
  annotations:
    # Enable cert-manager to create automatically the SSL certificate and store in Secret
    cert-manager.io/cluster-issuer: ca-issuer
    cert-manager.io/common-name: kibana.picluster.ricsanfre.com
spec:
  ingressClassName: nginx
  tls:
    - hosts:
        - kibana.picluster.ricsanfre.com
      secretName: kibana-tls
  rules:
    - host: kibana.picluster.ricsanfre.com
      http:
        paths:
          - path: /
            pathType: Prefix
            backend:
              service:
                name: efk-kb-http
                port:
                  number: 5601

Step 2: Apply manifest
```
kubectl apply -f manifest.yml
```
Step 3. Access to Kibana UI

UI can be access through http://kibana.picluster.ricsanfre.com using loging elastic and the password stored in <efk_cluster_name>-es-elastic-user.

Initial Kibana Setup (DataView configuration)

Kibana’s DataView must be configured in order to access Elasticsearch data.

Note: This configuration must be done once data from fluentd has been inserted in ES: A index (fluentd-<date>) containing data has been created.

Step 1: Open Kibana UI

Open a browser and go to Kibana’s URL (kibana.picluster.ricsanfre.com)
Step 2: Open “Management Menu”
Step 3: Select “Kibana - Data View” menu option and click on “Create data view”
Step 4: Set index pattern to fluentd-* and timestamp field to @timestamp and click on “Create Index”

Prometheus elasticsearh exporter installation

In order to monitor elasticsearch with prometheus, prometheus-elasticsearch-exporter need to be installed.

For doing the installation prometheus-elasticsearch-exporter official helm will be used.

Step 1: Add the prometheus community repository

helm repo add prometheus-community https://prometheus-community.github.io/helm-charts

Step 2: Fetch the latest charts from the repository
```
helm repo update
```

Step 3: Create values.yml for configuring the helm chart

---
# Elastic search password from secret
extraEnvSecrets:
  ES_USERNAME:
    secret: es-prometheus-user-file-realm
    key: username
  ES_PASSWORD:
    secret: es-prometheus-user-file-realm
    key: password


# Elastic search URI
es:
  uri: http://efk-es-http:9200

This config passes ElasticSearch API endpoint (uri) and the needed credentials through environement variables(ES_USERNAME and ES_PASSWORD). The es-prometheus-user-file-realm secret was created above when in Elasticsearch authentication

Step 3: Install prometheus-elasticsearh-exporter in the logging namespace with the overriden values

helm install -f values.yml prometheus-elasticsearch-exporter prometheus-community/prometheus-elasticsearch-exporter --namespace logging

When deployed, the exporter generates a Kubernetes Service exposing prometheus-elasticsearch-exporter metrics endpoint (/metrics on port 9108).

It can be tested with the following command:

curl prometheus-elasticsearch-exporter.logging.svc.cluster.local:9108/metrics
# HELP elasticsearch_breakers_estimated_size_bytes Estimated size in bytes of breaker
# TYPE elasticsearch_breakers_estimated_size_bytes gauge
elasticsearch_breakers_estimated_size_bytes{breaker="eql_sequence",cluster="efk",es_client_node="true",es_data_node="true",es_ingest_node="true",es_master_node="true",host="10.42.2.20",name="efk-es-default-0"} 0
elasticsearch_breakers_estimated_size_bytes{breaker="fielddata",cluster="efk",es_client_node="true",es_data_node="true",es_ingest_node="true",es_master_node="true",host="10.42.2.20",name="efk-es-default-0"} 0
elasticsearch_breakers_estimated_size_bytes{breaker="inflight_requests",cluster="efk",es_client_node="true",es_data_node="true",es_ingest_node="true",es_master_node="true",host="10.42.2.20",name="efk-es-default-0"} 0
elasticsearch_breakers_estimated_size_bytes{breaker="model_inference",cluster="efk",es_client_node="true",es_data_node="true",es_ingest_node="true",es_master_node="true",host="10.42.2.20",name="efk-es-default-0"} 0
...

Last Update: Oct 08, 2023

Comments:

On this page: