| Home | https://chris.funderburg.me/ |
| Github | https://github.com/bocan |
| https://www.linkedin.com/in/bocan/ |
| Home | https://chris.funderburg.me/ |
| Github | https://github.com/bocan |
| https://www.linkedin.com/in/bocan/ |
Next step in running my own Mastodon server:
✔️ Keep a fork of the Mastodon code in Github.
✔️ Sync that fork to my company's Gitlab server. (I own the company so it's ok.)
✔️ Add a CI/CD pipeline file that builds the Docker image and let it run each time the code changes.
✔️ Periodically review changes in Mastodon and sync that on Github to start the build.
Still to do : Get the mailgun account, then using Terraform, setup Cloud SQL Postgres, Redis, and maybe setup Elasticsearch
Building a Mastodon service - phase 1. Write some Terraform to spin up an autoscaling GKE cluster on Google Cloud, and then get the Terraform into my Gitlab with a CI/CD pipeline.
Further work will be to spin up a SQL Cloud Postgres instance, then Elasticsearch and Redis.
Bleh. Tedious work today. My client uses Varnish in their tech stack but they lost the ability to monitor it since they migrated to #Kubernetes. Varnish is like something from the dark ages and has no metrics API, so what I'm going to do is make cron run inside the Varnish container to export metrics to a file on shared storage - then run a Datadog agent in a sidecar to watch that file.
First bit's done so today's about getting the sidecar running.
New server, new #introduction.
My name's Chris. I'm a native Texan but I've lived in the United Kingdom for something like 27 years.
I currently run my own company as a consultant #DevOps engineer and #cloud architect focussing on #aws #gcp #kubernetes #terraform and #cicd. I've been deeply interested in tech most of my life. 🤓
Aside from tech, I've got an #allotment garden to grow flowers, veg, and fruit in; love to take pictures with my current #Canon R5, and help run a charity.
Moved over from mastodon.technology after it's sadly closing down.
That said, I'm thinking about building my own server - but I want it bullet-proof and highly available from day 1. That means running exclusively in Kubernetes, using a *cloud based database and Redis, etc. That's going to take a few weeks building it around my work hours.
* Probably going to use Google Cloud. I prefer AWS but GKE gives you one free Kubernetes cluster and you pay only for your worker nodes.