Logging

« list of all features

Applicable for:
ProviderAll Providers

Overview

Application and system logs can help you understand what is happening in your clusters. The logs are particularly useful for troubleshooting and monitoring cluster activity.

The Self-hosted ElasticSearch/Kibana or(and) OpenSearch/Dashboard can be used when you need dedicated Logging in the cluster.

NOTE You can find a lot of materials related to the logging feature usage, advanced customization and troubleshooting on the Kublr support portal. Find links and search tips in the “See also” section below.

For Cluster

You can select one of the options or both simultaneously:

  • Centralized Logging Agent (selected by default) - Centralized Logging transfers log records from all clusters created by the current Kublr platform to the platform-hosted ElasticSearch/OpenSearch.
  • Self-Hosted Logging - Self-Hosted Logging transfers cluster’s own log records to its self-hosted ElasticSearch/OpenSearch.

For the Self-Hosted Logging, you can:

  • Select between using Fluent Bit (default) and Fluentd.
  • Starting from Kublr version 1.26.0, use both Elasticsearch (selected by default) and OpenSearch. Note that you can enable both options but this may require additional cluster resources. Ensure that you have planned for additional resources accordingly or (preferably) select only one of the options.

ElasticSearch parameters

  • Number of Master nodes for ElasticSearch cluster. Note that the number of maste nodes should be 1, 3 or more. You can find more details here.
  • Number of Data nodes that hold data and perform data related operations such as CRUD, search, and aggregations.
  • Number of Client nodes that follow client requests.

The Persistence enabled option is selected by default.

Features-Logging

Optionally, you can also specify the Data node disk size in GB. Note that a 4 GB storage will be additionally created for each master node.

OpenSearch parameters

  • Count of master nodes - number of master replicas for OpenSearch cluster.
  • Count of data nodes - number of data replicas for OpenSearch cluster.
  • Data (data-hot) persistence size - size of persistence volume for “hot” data, default is 64Gi.
  • Hot Warm Architecture enabled - get details with “hot-warm-cold” search in the Kublr logging deployment options article and read this overview of the feature.

For Platform

Unlike cluster:

  • Logging cannot be disabled.
  • It has name Centralized Logging.
  • For AWS, it has an Add AWS CloudWatch Logs option.

Participate in Scenarios

  • Information will be provided soon.

See Also