I’m managing applications inside Kubernetes for more than 2 years for MySocialApp a social news feed solution and recently Referlab, an impressive referral marketing solution. If you follow me, you certainly know that I’ve made multiple Helm charts on distributed technologies like: Cassandra Helm chart Elasticsearch Helm chart Traefik Helm chart After several years of experience on it, you can trust me when I say managing statefulset on Kubernetes is not the easiest thing to do.
For my own company MySocialApp, I’m managing multiple Cassandra clusters on top of a Kubernetes on premise cluster. For those who never heard of this distributed database, here is the summary from the official website: The Apache Cassandra database is the right choice when you need scalability and high availability without compromising performance. Linear scalability and proven fault-tolerance on commodity hardware or cloud infrastructure make it the perfect platform for mission-critical data.
I’m using Kubernetes on an on premise cluster for MySocialApp. Today, I’m storing MySocialApp public images at Quay.io and I also wanted to store private images. I didn’t want to bootstrap a private registry for it to avoid maintaining it, having a distributed storage to maintain for it etc…but wanted a solution at a lower cost. I started to look at DockerHub and Quay.io. As far aas I saw, DockerHub do not provide private registry while Quay does.
Traefik (pronounced like traffic) is a modern HTTP reverse proxy and load balancer made to deploy microservices with ease. It supports several backends (Docker, Swarm, Kubernetes, Marathon, Mesos, Consul, Etcd, Zookeeper, BoltDB, Eureka, Amazon DynamoDB, Rest API, file…) to manage its configuration automatically and dynamically. In a previous post, I explained how to manually deploy it in HA mode. For MySocialApp (iOS and Android social app builder - SaaS), I had to automate it.
Traefik (pronounced like traffic) is a modern HTTP reverse proxy and load balancer made to deploy microservices with ease. It supports several backends (Docker, Swarm, Kubernetes, Marathon, Mesos, Consul, Etcd, Zookeeper, BoltDB, Eureka, Amazon DynamoDB, Rest API, file…) to manage its configuration automatically and dynamically. I wanted to deploy it on Kubernetes for its interesting features like: Connect to Kubernetes API to listen changes and perform on the fly updates Automatic SSL management through Let’s encrypt (SNI) Prometheus native integration HTTP/2 support I really like HAProxy, but in a Kubernetes case, it’s not the recommended solution because of its lake of features.
Kubernetes is an open-source system for automating deployment, scaling, and management of containerized applications. I’m using it for Nousmotards and decided to build an Ansible role for it to make it simpler to deploy. Why did I built it as other projects like Kargo already exists? I just wanted a simple and maintainable role based on CoreOS official documentation. That’s what I did: This role bootstrap a Kubernetes cluster based on CoreOS Container Linux for production usages.