Field notes from platform engineering.
Occasional writing on migrations, reliability practice, cost work, and the slow careful business of keeping cloud platforms running.
-
The 3-month AWS to OCI migration that taught me what cloud lock-in actually means
We moved 15 services, 3 MySQL databases, and 2 TB of data from AWS to Oracle Cloud in three months. The cutover was anticlimactic. The lessons that stuck were not.
-
Why we self-host our CI/CD
The decision narrative behind moving off managed runners onto GitHub ARC + ArgoCD. What we considered, what we rejected, what bit us in production, and what I'd tell someone weighing the same call today.
-
The Slack bot that runs our infrastructure
How we built a self-service platform inside Slack — API whitelisting, access requests, ephemeral environments — and why moving DevOps work into the channels people already use eliminated ~60% of recurring toil.