The build that kept dying at 2am
A deploy kept failing with no clear error. It turned out the 908 MB server was running out of memory mid-build. Here is how I spotted it and the one change that fixed it.
← Back homeLogs
Notes from the terminal.
Short write-ups on real problems and how I solved them. Full posts are coming soon.
Moving a live site with almost no downtime
Swapping servers without taking a site offline is mostly about doing DNS and SSL in the right order. My checklist for a clean cutover.
Why I keep Terraform state in S3
Storing infrastructure state on your laptop is a trap. A short note on remote state, locking, and why your team will thank you.
Deploys should be boring
If a deploy makes you nervous, something is wrong. A few habits that turn releases into a non-event.