Introduction
An OpenStack cloud generates a large quantity of log messages. By default,
there’s no way to visually see a Private Cloud’s log messages. To diagnose
issues in logs requires using SSH and grep, which can be cumbersome due to
the number of hosts and number of OpenStack services. Elasticsearch and Kibana
(ELK) could be leveraged to see all of your cloud’s log files from a single
location in your browser. This feature set is not enabled by default. In this
guide we walk you through how to enable the ELK stack for your Private Cloud
using Kolla Ansible.
Prerequisites
Prepare Kolla Ansible
This guide explains how to configure your cloud using Kolla Ansible. Any time you work with Kolla Ansible, you must prepare a shell environment. For more, see How to Prepare and Use Kolla Ansible.
All commands are to be executed from the control plane node in which Kolla Ansible has been prepared.
Root Access to OpenStack Control Plane
Root access to your cloud’s control plane nodes is required.
How to Enable Central Logging
To enable the ELK stack, in /etc/kolla/globals.yml ensure the following is
set:
enable_central_logging: 'yes'
To enforce data retention policies, enable Elasticseach Curator with:
enable_elasticsearch_curator: 'yes'
Note! — Enabling Elasticseach Curator can help prevent your cloud’s local disks from filling up by enforcing retention policies for Elasticsearch data.
Kolla Ansible generates a default configuration for Elasticsearch Curator
which can later be found in the elasticsearch_curator Docker container as
elasticsearch_curator:/etc/elasticsearch-curator/actions.yml.
Next, to deploy the configuration changes, use:
# kolla-ansible -i /etc/fm-deploy/kolla-ansible-inventory reconfigure
Prevent Root Disk from Filling
You can enable Elasticsearch Curator to enforce disk retention polices to prevent your cloud’s disks from filling up.
Reference
Kolla Ansible’s Central Logging guide.