Why Kubernetes and Terraform are the pillars of modern IT infrastructure

• 16.10.2018
• DevOps
• 0

In the majority of Spaceport projects, we have to design, deploy and maintain the underlying IT infrastructure for a certain product or service. Terraform and Kubernetes are the tools for that.

Terraform was developed by Hashicorp, the company behind Vagrant, Consul, Vault, Nomad and other tools from HashiStack. This is a configuration orchestration tool that is incredible for provisioning, adjusting and destroying the virtual server environments. It is available both as an DevOps-as-a-Service enterprise-grade offer from Hashicorp and as an open-source solution able to work with a variety of Cloud Service Providers to create multi-cloud ecosystems.

If you worked with AWS CloudFormation, just imagine a similar system that works with any set of components from any number of cloud computing providers — and you will see why Terraform rocks. Terraform manifests work equally well on AWS, Azure, GCP, DigitalOcean, VMware, allow spawning fully-functional environments in literally no time, and enable automated setup of all the needed software and libraries, DNS and networking configuration, as well as all the required virtual machines.

One might mention that there are multiple configuration management tools like Chef, Ansible, Salt, or Puppet that can perform these tasks — and one would be right. We would say even more, Ansible playbooks are great for certain aspects of infrastructure configuration, complete with installing the required versions of infrastructure components.

However, the core benefit of Terraform manifests is that they are written in descriptive, not procedural language and are cloud-agnostic. This means that any needed state of the infrastructure can be codified and versioned like any other code through Github. Any operation output can become the input for any other operation, allowing managing the immutable infrastructure as code — the state unreachable by older generation DevOps tools.

Containers should be managed using Kubernetes

Indtroduction of Docker containers simply revolutionized the way the IT industry develops, deploys and maintains the software. When your app has everything it needs to run packed into a neat code container, can run on any underlying infrastructure and any required number of containers can be launched with ease — it’s miles higher than the earlier ways of manual server configuration and app setup.

There were several attempts to create a tool for efficient management of containerized apps. You might have heard about Docker Swarm or Mesos/Marathon container management systems.

Kubernetes is an open-source container management platform originally developed by Google and included by default into many services of Google Cloud Platform, AWS, and MS Azure. This multi-layered tool implements the Infrastructure as Code principle of DevOps, allowing to manage each infrastructure layer — from a single container to pods, nodes, namespaces and clusters, along with networking and physical hosts — independently. Why?

Because it allows the system to be fail-proof. If an error occurs, a monitoring agent forms a report and initiates the reprovisioning of the faulty element, so there are no major crashes. With Kubernetes, it’s easier to replace the faulty instance on the move than to stop and reboot the whole system. This is invaluable as any DevOps engineer will tell you.

Conclusions on the role of Kubernetes and Terraform

Using Kubernetes and Terraform manifests allows treating the IT infrastructure states as the code. These documents can be stored on GitHub, versioned and automatically used in software delivery pipelines. Every developer can fork the needed infrastructure state for a project, adjust the required parameters and build the needed environment in mere minutes. This is stellar years ahead of Waterfall software development with endless back-and-forth bounces of requests between the Devs and the Ops teams.

Using this DevOps approach allows shortening the product time-to-market by no less than 50%, as the recent research on the State of DevOps adoption shows. There are numerous other benefits, like increased disaster recovery time, fewer release issues, dramatically decreased infrastructure expenses, etc.

The point is — using these capabilities to the full extent is possible only if you cooperate with a trustworthy DevOps team, like Spaceport. We can execute the projects you have at hand and help your in-house IT team leverage the new solutions with maximum efficiency. Contact Spaceport team, we are always glad to assist!