A newer version is available. For the latest information, see the current release documentation.
« Quickstart Securing Logstash API »
Elastic Docs ›Elastic Cloud on Kubernetes [2.12] ›Orchestrating Elastic Stack applications ›Run Logstash on ECK

Configuration

edit

Configuration

edit

This functionality is in technical preview and may be changed or removed in a future release. Elastic will work to fix any issues, but features in technical preview are not subject to the support SLA of official GA features.

Upgrade the Logstash specification

edit

You can upgrade the Logstash version or change settings by editing the YAML specification. ECK applies the changes by performing a rolling restart of Logstash Pods.

Logstash configuration

edit

Define the Logstash configuration (the ECK equivalent to logstash.yml) in the spec.config section:

apiVersion: logstash.k8s.elastic.co/v1alpha1
kind: Logstash
metadata:
  name: quickstart
spec:
  version: 8.16.0
  count: 1
  elasticsearchRefs:
  - name: quickstart
    clusterName: qs
  config: 
    pipeline.workers: 4
    log.level: debug

Customize Logstash configuration using logstash.yml settings here

Alternatively, you can provide the configuration through a Secret specified in the spec.configRef section. The Secret must have an logstash.yml entry with these settings:

apiVersion: logstash.k8s.elastic.co/v1alpha1
kind: Logstash
metadata:
  name: quickstart
spec:
  version: 8.16.0
  count: 1
  elasticsearchRefs:
  - name: quickstart
    clusterName: qs
  configRef:
    secretName: quickstart-config
---
apiVersion: v1
kind: Secret
metadata:
  name: quickstart-config
stringData:
  logstash.yml: |-
    pipeline.workers: 4
    log.level: debug

Configuring Logstash pipelines

edit

Define Logstash pipelines in the spec.pipelines section (the ECK equivalent to pipelines.yml):

apiVersion: logstash.k8s.elastic.co/v1alpha1
kind: Logstash
metadata:
  name: quickstart
spec:
  version: 8.16.0
  count: 1
  elasticsearchRefs:
    - clusterName: qs
      name: quickstart
  pipelines:
    - pipeline.id: main
      config.string: |
        input {
          beats {
            port => 5044
          }
        }
        output {
          elasticsearch {
            hosts => [ "${QS_ES_HOSTS}" ]
            user => "${QS_ES_USER}"
            password => "${QS_ES_PASSWORD}"
            ssl_certificate_authorities => "${QS_ES_SSL_CERTIFICATE_AUTHORITY}"
          }
        }

Alternatively, you can provide the pipeline configuration through a Secret specified in the spec.pipelinesRef element. The Secret must have a logstash.yml entry with this configuration:

apiVersion: logstash.k8s.elastic.co/v1alpha1
kind: Logstash
metadata:
  name: quickstart
spec:
  version: 8.16.0
  count: 1
  elasticsearchRefs:
    - clusterName: qs
      name: quickstart
  pipelinesRef:
    secretName: quickstart-pipeline
---
apiVersion: v1
kind: Secret
metadata:
  name: quickstart-pipeline
stringData:
  pipelines.yml: |-
    - pipeline.id: main
      config.string: |
        input {
          beats {
            port => 5044
          }
        }
        output {
          elasticsearch {
            hosts => [ "${QS_ES_HOSTS}" ]
            user => "${QS_ES_USER}"
            password => "${QS_ES_PASSWORD}"
            ssl_certificate_authorities => "${QS_ES_SSL_CERTIFICATE_AUTHORITY}"
          }
        }

Logstash on ECK supports all options present in pipelines.yml, including settings to update the number of workers, and the size of the batch that the pipeline will process. This also includes using path.config to point to volumes mounted on the Logstash container:

apiVersion: logstash.k8s.elastic.co/v1alpha1
kind: Logstash
metadata:
  name: quickstart
spec:
  version: 8.16.0
  count: 1
  elasticsearchRefs:
    - clusterName: qs
      name: quickstart
  pipelines:
    - pipeline.id: main
      config.string: |
        input {
          beats {
            port => 5044
          }
        }
        output {
          elasticsearch {
            hosts => [ "${QS_ES_HOSTS}" ]
            user => "${QS_ES_USER}"
            password => "${QS_ES_PASSWORD}"
            ssl_certificate_authorities => "${QS_ES_SSL_CERTIFICATE_AUTHORITY}"
          }
        }

Logstash persistent queues (PQs) and dead letter queues (DLQs) are not currently managed by the Logstash operator, and using them will require you to create and manage your own Volumes and VolumeMounts

Defining data volumes for Logstash

edit

[2.9.0] Added in 2.9.0.

Volume support for Logstash is a breaking change to earlier versions of ECK and requires you to recreate your Logstash resources.

Specifying the volume claim settings

edit

A PersistentVolume called logstash-data is created by default. It maps to /usr/share/logstash/data for persistent storage, which is typically used for storage from plugins.

By default, the logstash-data volume claim is a 1.5Gi volume, using the standard StorageClass of your Kubernetes cluster. You can override the default by adding a spec.volumeClaimTemplate section named logstash-data.

For production workloads, you should define your own volume claim template with the desired storage capacity and (optionally) the Kubernetes storage class to associate with the persistent volume. To override this volume claim for data usages, the name of this volume claim must be logstash-data.

This example updates the default data template to increase the storage to 2Gi for the Logstash data folder:

apiVersion: logstash.k8s.elastic.co/v1alpha1
kind: Logstash
metadata:
  name: logstash
spec:
  # some configuration attributes omitted for brevity here
  volumeClaimTemplates:
    - metadata:
        name: logstash-data # Do not change this name unless you set up a volume mount for the data path.
      spec:
        accessModes:
          - ReadWriteOnce
        resources:
          requests:
            storage: 2Gi

The default volume size will likely be insufficient for production workloads, especially when you are using:

  • the persistent queue (PQ) feature
  • dead letter queues (DLQ), or
  • Logstash plugins that make heavy use of temporary storage.

Increase the storage capacity, or consider creating separate volumes for these use cases.

You can add separate storage by including an additional spec.volumeClaimTemplate along with a corresponding spec.podTemplate.spec.containers.volumeMount for each requested volume.

This example shows how to setup separate storage for a PQ:

apiVersion: logstash.k8s.elastic.co/v1alpha1
kind: Logstash
metadata:
  name: logstash
spec:
  # some configuration attributes omitted for brevity here
  volumeClaimTemplates:
    - metadata:
        name: pq 
      spec:
        accessModes:
        - ReadWriteOnce
        resources:
          requests:
            storage: 10Gi
  podTemplate:
    spec:
      containers:
      - name: logstash
        volumeMounts:
        - mountPath: /usr/share/logstash/pq 
          name: pq  
          readOnly: false
  config:
    log.level: info
    queue.type: persisted
    path.queue: /usr/share/logstash/pq

The name values in the volumeMount for the container in the podTemplate section and the name of the volumeClaimTemplate must match.

Set the path.queue setting in the configuration to match the mountPath in the volumeMount.

This example shows how to configure Logstash with a Dead Letter Queue setup on the main pipeline, and a separate pipeline to read items from the DLQ.

apiVersion: logstash.k8s.elastic.co/v1alpha1
kind: Logstash
metadata:
  name: logstash
spec:
   # some configuration attributes omitted for brevity here
   podTemplate:
    spec:
      containers:
      - name: logstash
        volumeMounts:
        - mountPath: /usr/share/logstash/dlq 
          name: dlq  
          readOnly: false
  volumeClaimTemplates:
    - metadata:
      name: dlq 
      spec:
        accessModes:
        - ReadWriteOnce
        resources:
          requests:
            storage: 10Gi
  pipelines:
    - pipeline.id: beats
      dead_letter_queue.enable: true
      path.dead_letter_queue: /usr/share/logstash/dlq 
      config.string: |
        input {
          beats {
            port => 5044
          }
        }
        output {
          elasticsearch {
            hosts => [ "${ECK_ES_HOSTS}" ]
            user => "${ECK_ES_USER}"
            password => "${ECK_ES_PASSWORD}"
            ssl_certificate_authorities => "${ECK_ES_SSL_CERTIFICATE_AUTHORITY}"
          }
        }
    - pipeline.id: dlq_read
      dead_letter_queue.enable: false
      config.string: |
        input {
          dead_letter_queue {
            path => "/usr/share/logstash/dlq" 
            commit_offsets => true
            pipeline_id => "beats"
            clean_consumed => true
          }
        }
        filter {
          mutate {
            remove_field => "[geoip][location]"
          }
        }
        output {
          elasticsearch {
            hosts => [ "${ECK_ES_HOSTS}" ]
            user => "${ECK_ES_USER}"
            password => "${ECK_ES_PASSWORD}"
            ssl_certificate_authorities => "${ECK_ES_SSL_CERTIFICATE_AUTHORITY}"
          }
        }

The name values in the volumeMount for the container in the podTemplate section and the name of the volumeClaimTemplate must match.

Set the path.dead_letter_queue setting in the pipeline config to match the mountPath in the volumeMount for pipelines that are writing to the Dead Letter Queue, and set the path setting of the dead_letter_queue plugin for the pipeline that will read from the Dead Letter Queue.

Updating the volume claim settings

edit

If the storage class allows volume expansion, you can increase the storage requests size in spec.volumeClaimTemplates. ECK updates the existing PersistentVolumeClaims accordingly, and recreates the StatefulSet automatically.

If the volume driver supports ExpandInUsePersistentVolumes, the filesystem is resized online. In this case, you do not need to restart the Logstash process or re-create the Pods.

If the volume driver does not support ExpandInUsePersistentVolumes, you must manually delete Pods after the resize so that they can be recreated automatically with the expanded filesystem.

Any other changes in the volumeClaimTemplates—​such as changing the storage class or decreasing the volume size—​are not allowed. To make changes such as these, you must fully delete the Logstash resource, delete and recreate or resize the volume, and create a new Logstash resource.

Before you delete a persistent queue (PQ) volume, ensure that the queue is empty. When using the PQ, we recommend setting queue.drain: true on the Logstash Pods to ensure that the queue is drained when Pods are shutdown. Note that you should also increase the terminationGracePeriodSeconds to a large enough value to allow the queue to drain.

This example shows how to configure a Logstash resource to drain the queue and increase the termination grace period.

apiVersion: logstash.k8s.elastic.co/v1alpha1
kind: Logstash
metadata:
  name: logstash
spec:
  # some configuration attributes omitted for brevity here
  config:
    queue.drain: true
  podTemplate:
    spec:
      terminationGracePeriodSeconds: 604800

A Kubernetes known issue: Kubernetes may not honor terminationGracePeriodSeconds settings greater than 600. A queue of a terminated Pod may not be fully drained, even when queue.drain: true is set and a high terminationGracePeriodSeconds is configured.

In this technical preview, there is currently no way to drain a dead letter queue (DLQ) automatically before Logstash shuts down. To manually drain the queue, first stop sending data to it, by either disabling the DLQ feature, or disabling any pipelines that send to a DLQ. Then wait for events to stop flowing through any pipelines reading from the input.

EmptyDir

edit

If you are not concerned about data loss, you can use an emptyDir volume for Logstash data.

The use of emptyDir in a production environment may cause permanent data loss. Do not use with persistent queues (PQs), dead letter queues (DLQs), or with any plugin that requires persistent storage to keep track of state between restarts of Logstash.

Plugins that require persistent storage include any plugin that stores state locally. These plugins typically have a configuration parameter that includes the name path or directory, not including paths to static content, such as certificates or keystores. Examples include the sincedb_path setting for the file, dead_letter_queue and s3 inputs, the last_run_metadata_path for the JDBC input, aggregate_maps_path for the aggregate filter, and temporary_directory for the s3 output, used to aggregate content before uploading to s3.

spec:
  count: 5
  podTemplate:
    spec:
      volumes:
      - name: logstash-data
        emptyDir: {}

Using Elasticsearch in Logstash pipelines

edit

The spec.elasticsearchRefs section provides a mechanism to help configure Logstash to establish a secured connection to one or more ECK managed Elasticsearch clusters. By default, each elasticsearchRef will target all nodes in its referenced Elasticsearch cluster. If you want to direct traffic to specific nodes of your Elasticsearch cluster, refer to Traffic Splitting for more information and examples.

When you use elasticsearchRefs in a Logstash pipeline, the Logstash operator creates the necessary resources from the associated Elasticsearch cluster, and provides environment variables to allow these resources to be accessed from the pipeline configuration. Environment variables are replaced at runtime with the appropriate values. The environment variables have a fixed naming convention:

  • NORMALIZED_CLUSTERNAME_ES_HOSTS
  • NORMALIZED_CLUSTERNAME_ES_USER
  • NORMALIZED_CLUSTERNAME_ES_PASSWORD
  • NORMALIZED_CLUSTERNAME_ES_SSL_CERTIFICATE_AUTHORITY

where NORMALIZED_CLUSTERNAME is the value taken from the clusterName field of the elasticsearchRef property, capitalized, and - transformed to _ - eg, prod-es, would becomed PROD_ES.

The clusterName value should be unique across all referenced Elasticsearches in the same Logstash spec.

The Logstash ECK operator creates a user called eck_logstash_user_role when an elasticsearchRef is specified. This user has the following permissions:

  "cluster": ["monitor", "manage_ilm", "read_ilm", "manage_logstash_pipelines", "manage_index_templates", "cluster:admin/ingest/pipeline/get",],
  "indices": [
    {
      "names": [ "logstash", "logstash-*", "ecs-logstash", "ecs-logstash-*", "logs-*", "metrics-*", "synthetics-*", "traces-*" ],
      "privileges": ["manage", "write", "create_index", "read", "view_index_metadata"]
    }

You can update user permissions to include more indices if the Elasticsearch plugin is expected to use indices other than the default. Check out Logstash configuration with a custom index sample configuration that creates a user that writes to a custom index.

This example demonstrates how to create a Logstash deployment that connects to different Elasticsearch instances, one of which is in a separate namespace:

apiVersion: logstash.k8s.elastic.co/v1alpha1
kind: Logstash
metadata:
  name: quickstart
spec:
  version: 8.16.0
  count: 1
  elasticsearchRefs:        
    - clusterName: prod-es  
      name: prod
    - clusterName: qa-es    
      name: qa
      namespace: qa
  pipelines:
    - pipeline.id: main
      config.string: |
        input {
          beats {
            port => 5044
          }
        }
        output {
          elasticsearch {   
            hosts => [ "${PROD_ES_ES_HOSTS}" ]
            user => "${PROD_ES_ES_USER}"
            password => "${PROD_ES_ES_PASSWORD}"
            ssl_certificate_authorities => "${PROD_ES_ES_SSL_CERTIFICATE_AUTHORITY}"
          }
          elasticsearch {   
            hosts => [ "${QA_ES_ES_HOSTS}" ]
            user => "${QA_ES_ES_USER}"
            password => "${QA_ES_ES_PASSWORD}"
            ssl_certificate_authorities => "${QA_ES_ES_SSL_CERTIFICATE_AUTHORITY}"
          }
        }

Define Elasticsearch references in the CRD. This will create the appropriate Secrets to store certificate details and the rest of the connection information, and create environment variables to allow them to be referred to in Logstash pipeline configurations.

This refers to an Elasticsearch cluster residing in the same namespace as the Logstash instances.

This refers to an Elasticsearch cluster residing in a different namespace to the Logstash instances.

Elasticsearch output definitions - use the environment variables created by the Logstash operator when specifying an ElasticsearchRef. Note the use of "normalized" versions of the clusterName in the environment variables used to populate the relevant fields.

Connect to an external Elasticsearch cluster

edit

Logstash can connect to external Elasticsearch cluster that is not managed by ECK. You can reference a Secret instead of an Elasticsearch cluster in the elasticsearchRefs section through the secretName attribute:

apiVersion: v1
kind: Secret
metadata:
  name: external-es-ref
stringData:
  url: https://abcd-42.xyz.elastic-cloud.com:443 
  username: logstash_user 
  password: REDACTED 
  ca.crt: REDACTED 
---
apiVersion: logstash.k8s.elastic.co/v1alpha1
kind: Logstash
metadata:
  name: quickstart
spec:
  version: 8.16.0
  count: 1
  elasticsearchRefs:
    - clusterName: prod-es
      secretName: external-es-ref 
  monitoring:
    metrics:
      elasticsearchRefs:
      - secretName: external-es-ref 
    logs:
      elasticsearchRefs:
      - secretName: external-es-ref

The URL to reach the Elasticsearch cluster.

The username of the user to be authenticated to the Elasticsearch cluster.

The password of the user to be authenticated to the Elasticsearch cluster.

The CA certificate in PEM format to secure communication to the Elasticsearch cluster (optional).

The secretName and name attributes are mutually exclusive, you have to choose one or the other.

Please always specify the port in URL when connecting to an external Elasticsearch Cluster.

Expose services

edit

By default, the Logstash operator creates a headless Service for the metrics endpoint to enable metric collection by the Metricbeat sidecar for Stack Monitoring:

kubectl get service quickstart-ls-api
NAME                TYPE        CLUSTER-IP   EXTERNAL-IP   PORT(S)    AGE
quickstart-ls-api   ClusterIP   None         <none>        9600/TCP   48s

Additional services can be added in the spec.services section of the resource:

services:
  - name: beats
    service:
      spec:
        ports:
        - port: 5044
          name: "winlogbeat"
          protocol: TCP
        - port: 5045
          name: "filebeat"
          protocol: TCP

Pod configuration

edit

You can customize the Logstash Pod using a Pod template, defined in the spec.podTemplate section of the configuration.

This example demonstrates how to create a Logstash deployment with increased heap size and resource limits.

apiVersion: logstash.k8s.elastic.co/v1alpha1
kind: Logstash
metadata:
  name: logstash-sample
spec:
  version: 8.16.0
  count: 1
  elasticsearchRefs:
    - name: "elasticsearch-sample"
      clusterName: "sample"
  podTemplate:
    spec:
      containers:
      - name: logtash
        env:
        - name: LS_JAVA_OPTS
          value: "-Xmx2g -Xms2g"
        resources:
          requests:
            memory: 1Gi
            cpu: 0.5
          limits:
            memory: 4Gi
            cpu: 2

The name of the container in the Pod template must be logstash.

« Quickstart Securing Logstash API »