Recently I’ve been looking at how to configure EC2 autoscaling schedules for EKS implementations, specifically delivering these schedule configurations via Terraform. This sounds like it should be rather simple on the surface but after getting the initial configuration to work an issue of idempotency presents itself. In this post I want to look at the issues presented and how to overcome them. Autoscaling Groups and Schedules When an managed EKS . . .
An often required feature of any declarative software or scripting is to work with time values. Much of the time this requirement doesn’t crawl out of the woodwork until you’ve been working with it for a while (at least that’s usually my experience). It was a relief to learn that Terraform does have this function, but the use is a little out of the ordinary and takes a bit of . . .
In a previous post we looked at setting up centralised Terraform state management using S3 for AWS provisioning (as well as using Azure Object Storage for the same solution in Azure before that). What our S3 solution lacked however is a means to achieve State Locking, I.E. any method to prevent two operators or systems from writing to a state at the same time and thus running the risk of . . .
In previous posts we’ve looked at how to look up Secrets from Hashicorp Vault using Ansible and Ansible Tower. We’ve also taken a look at how to integrate Azure Key Vault with Ansible Tower, however I’ve never gotten round to taking a look at how to integrate Ansible itself with Azure Key Vault (without the use of Tower). Whilst I’ve largley moved away from using Azure Key Vault in favour . . .
Recently I’ve spent a good amount of time looking at options for managing Kubernetes Secrets with Vault. Hashicorp being a great supporter of the Cloud Native philosophy, it’s little surprise to find that they provide a multitude of options to integrate with Kubernetes and provide extensive documentation here. for my needs I found that the suggested configurations were either unsuitable or required a degree of over-engineering so I’m going to . . .
In a previous post we looked at a method to use Terraform’s output function to export return data and load it in to an external YAML file for consumption by Ansible. While this is a useful function it’s a little topheavy, and if we just want to pass data in to another Terraform configuration in order to run an apply operation, we have a means to work a lot more . . .
UPDATED 11/2020: Have a look at a different method for this configuration better suited to CI/CD. In a previous post we looked at how to use Terraform provision and authenticate with Clusters using AWS’ Elastic Kubernetes Service (EKS) using the somewhat unique authentication method of it’s webhook token method leveraging aws-iam-authenticator. Once we get past that point however we still have another permission hurdle to overcome, specifically how we handle . . .
NOTE: The sample code used here is hosted in my GitHub here. Recently I’ve been getting my hands dirtier and dirtier with Kubernetes but there’s some interesting oddities that only occur in Elastic Kubernetes Service (EKS), the AWS PaaS Kubernetes platform, especially when it comes to how you can authenticate. As Kubernetes is strongly driven by a declarative (and by extension Infrastructure as Code) philosophy, it makes perfect sense that . . .
In the last post we looked at how to automate the creation of GKE Kubernetes clusters in GCP, however the deployment of workloads to these clusters was still something of a manual process. Enter Helm; the a package manager for Kubernetes. If this is your first time hearing those words it can be pretty confusing, typically when we think of package management we think of application packages or application dependencies . . .
Google Cloud Platform tends to be forgotten from the conversation a lot when talking about public cloud offerings, however their hosted Kubernetes offering GKE (Google Kubernetes Engine) has for me been the best of the major offerings for getting to grips with the platform and the best reason to use GCP at all. Without much issue we can get Terraform integrated with GCP, provision and scale out clusters as we . . .