terraform-kibana-dev

Deploy any branch from a Kibana repository to AWS, Azure, or GCP for testing in development.

This is intended for development purposes only; it will not run Kibana in a production environment.

Overview

The config in each directory performs the same basic steps in each cloud environment:

• Creates a single VM running Ubuntu 18.04 (LTS)
• Clones the Kibana repo and checks out the branch indicated in variables.tf or terraform.tfvars
• Starts an Elasticsearch dev server in the background via yarn es snapshot
• Runs Kibana in dev mode on the default port indicated in kibana.dev.yml (5601), unless configured to do so differently
• Creates a firewall to allow inbound traffic on Kibana's port and allow SSH access to the instance via the public key provided in variables.tf or terraform.tfvars
• Writes logs to ~/kibana/kibana.log

Usage

Refer to the provider-specific READMEs in each directory for usage instructions.

Requirements

• Terraform 0.14.0 or higher
• Linux or macOS (I haven't tested this on Windows)
• Accounts with each cloud provider you wish to use
• Cloud provider service credentials (see each README for details)

Reminders

Don't use for production deployments.

This does not create a production environment. It runs Elasticsearch as a single-node cluster on the same instance as Kibana, and no data is persisted between restarts.

Be patient.

Running terraform apply takes awhile as it needs to not only provision the instances, but also run kbn bootstrap, which can take several minutes.

While you should be able to SSH into the instance as soon as apply has completed, you won't be able to see Kibana in the browser right away. This is because the Kibana server is kicked off in the background, and needs to build all of the plugins and run optimizer to bundle the client-side code before serving traffic. This process typically takes 5-10 minutes to finish after apply has completed.

The code is free, but running the VMs isn't.

Kibana needs more power than each cloud provider's free tier can offer, so this will cost a small amount of money to run. Don't forget to terraform destroy when you are done testing. I tried my best to pick the cheapest suitable instances I could find in each us-west region, but you can also configure instance sizes & regions via terraform.tfvars should you wish to do so. With the default settings, costs are in the range of $0.10-$0.30 USD per hour at the time of this writing.