About

I’m Galust, a DevOps engineer. Right now I run production Kubernetes for a fintech lending group: managed Kubernetes in the cloud, VMware and our own hardware, plus all the stateful pieces that come with money — PostgreSQL, RabbitMQ, Redis, Redpanda and Vault.

The part of the job I value most is what happens when things break. In October 2026 our cloud provider lost two of its three availability zones within a few days. Production went down, the surviving worker nodes had no internet access, and everything had to be pulled out and started somewhere else. I backed up the clusters with Velero running outside the cluster, saved the images and the state of Vault, Redis, RabbitMQ and Redpanda, and then restored them on new Kubernetes in our office within a day. The backup part is described in this post.

How I work

• Evidence before guesses. Audit logs, Helm history, log offsets and metrics usually answer “who did this” and “where did the data go” faster than theories.
• Dry run first. Anything that touches production in bulk gets a dry run, check-and-set writes and a rollback log.
• Write it down. After an incident I leave a runbook, so the next person doesn’t start from zero.

Teaching

I’m a senior curator at ProstoDevops, an online DevOps school. I look after about a hundred students, run reviews and mock interviews, and build lessons and hands-on labs on Linux, networking, Docker, Kubernetes, Terraform, Ansible and Argo CD. Explaining something simply is the best check that I understand it myself.

Contact

I speak Russian, Armenian and English. The fastest way to reach me is Telegram; email works too: galust.arakelyan25@gmail.com. My CV is here.