elliotsegler / elk6

Elasticsearch, Logstash, Kibana stack for version 6

ELK6 stack

This project is an amazon-hosted ELK stack for version 6. It uses the existing AWS ElasticSearch service and creates nodes for Kibana and Logstash.

How to use it

Requirements

• An AWS account (duh!), with at least one VPC configured
• packer (https://www.packer.io/) on your local machine
• Ruby 2.x on your local machine. Macs should have this by default, unless they are really old, if you're on Linux and you don't already have it then it should be simple enough to install from your favourite package management system
• AWS credentials for Packer, normally using environment variables or profiles

Step one - build the images

• Check out this repo to your local machine
• Make sure that you set up credentials for your AWS account in a Terminal session
• Then you should be able to build the Logstash image by running utils/build-image.rb logstash from the repo directory
• Make a note of the AMI ID that it has created
• Now build the Kibana image by running utils/build-image.rb kibana from the repo directory
• Again, make a note of the AMU ID that it has created
• Note - don't try to build the two in parallel, the build-image script creates a tempfile for packer that may get over-written if you try

Step two - deploy the stack

• Log in to the AWS Console and go to Cloud Formation
• Click Create Stack and upload cloudformation/fullstack.yaml from the repo
• Most of the parameters should be self-explanatory. Note - you must specify exactly two subnets to deploy into, because of how Amazon's ElasticSearch service works.
• Paste in the AMI IDs of the Logstash and Kibana images that you created

Step two A - tweaks

• In order to allow Logstash to talk directly to ElasticSearch you must manually set a policy on your created ElasticSearch domain to allow nodes to access without IAM authentication.
• In the AWS Console go to Elasticsearch Service, then select the Domain that CloudFormation created for you
• Click 'Modify Access Policy' and then select 'Do Not Require Signing with an IAM Credential' from the dropdown list
• Click Submit
• Wait for the change to take effect

This means that only the hosts identified in the Security Groups for the domain can access it. The Cloudformation sets that up as the created Logstash and Kibana nodes only, anything else will be firewalled out. I'd also recommend deploying this in a private VPC to minimise the attack surface.

Step three - use it

• Install Filebeat or similar onto the nodes you want to monitor - see https://www.elastic.co/products/beats for more information
• Use the Logstash endpoint as the target for Filebeat etc. to push to
• Set up a DNS CNAME to get a more usable Kibana endpoint than the auto-generated one
• The Logstash and Kibana endpoints are shown as CloudFormation outputs and they are exported. You can either use these directly in your Beats configs (not recommended) or import them as external references into another CloudFormation stack (see https://docs.aws.amazon.com/AWSCloudFormation/latest/UserGuide/walkthrough-crossstackref.html) and pass that value into your Beats config.

Logstash filter configurations

The deployment and testing of filter configurations needs some work. At present, any files that are present in config/logstash/conf.d will be placed into /etc/logstash/conf.d as the logstash image is built, with no testing.

Syntax errors will result in logstash failing to start up and constant restarts from the autoscaling group.

This will be improved in a future PR...

Instance details and security

The Logstash and Kibana instances are built from Ubuntu 16.04 LTS by default (but it's simple to change the source AMI in the packer-common.yaml configuration). A full update is run during the image build process, so to keep up to date with security patches it's recommended to rebuild your instances regularly.