Tuning Logstash pipelines for cost, not throughput.
For a year I optimized our pipelines for events-per-second. It took a small surprise on the AWS invoice to realize we were measuring the wrong thing entirely.
For a year I optimized our pipelines for events-per-second. It took a small surprise on the AWS invoice to realize we were measuring the wrong thing entirely.
A small accounting of every alarm I created, used, and forgot — and what they collectively did to a quarterly bill.
What worked, what didn't, and the one thing I'd absolutely do again on the next project.
Throughput is a vanity metric. The bill is the reality.
A walkthrough of the smallest reliable single-node setup I'd actually run in production.
A small confession, a small rewrite, and the case for switching to Python earlier than you think.
An on-call survival kit that will not solve your problem, but will tell you who to wake up next.
Five rules I keep coming back to after maintaining the same playbooks for three years.
In which cron is fine, systemd timers are also fine, and the choice matters less than you think.
Things I'd do differently if I were starting another small CLI tomorrow.
Composition over configuration, with examples from a year of slow rewrites.
Yes, I've seen the new terminals. I'm still here.
A practical writeup of the Ansible bot, focusing on the small parts that turned out to matter.
A boring topic, made slightly less boring by an embarrassing $$$ mistake I made first.
A short note on starting this notebook, and the kinds of writing I want to put here.