When production is down, users don't care why it happened - they care that it's working again. Roll back the deployment, disable the feature, or reroute traffic first. Once things are stable, investigate what actually went wrong.

The fastest way to make an outage worse is changing something you don't understand. Spend a few minutes checking logs, metrics, and recent deployments before running commands. A little patience usually saves hours of recovery.

Whenever possible, change one thing at a time. Small deployments are easier to test, easier to rollback, and much easier to debug when something breaks.

The first thing you think is broken usually isn't. Always ask yourself, "What evidence do I have?" Verify your assumptions before acting.

Guessing is a sign that your monitoring isn't telling you enough. Before trying random fixes, collect more information. Good decisions come from good visibility.

If you've done the same task three or four times, automate it. Every manual step is another chance for mistakes, especially during stressful incidents.

You'll forget today's solution in six months. Future teammates won't know it either. Good documentation saves far more time than it takes to write.

Don't try to master ten monitoring tools at once. Become excellent with the one your team uses every day. Deep knowledge is far more valuable than surface-level familiarity with everything.

After every production issue, ask yourself one question: "What will I do differently next time?" Small lessons compound into experience.

If something feels annoying today, it'll be unbearable six months from now. Leave systems cleaner than you found them. Your future self will thank you.

If an alert doesn't require immediate action, it shouldn't page anyone. Too many noisy alerts train engineers to ignore the important ones.

Rolling back isn't admitting defeat - it's protecting users. You can always investigate later, but you can't recover lost customer trust.

Silence makes people assume the worst. During incidents, keep sharing updates, even if the update is simply "We're still investigating."

If your work might affect someone else, let them know early. Proactive communication builds trust and prevents surprises.

Nobody expects juniors to know everything. Asking a five-minute question is much better than spending two hours fixing a mistake that could have been avoided.

Fixing production is only half the job. Follow through with documentation, cleanup, and preventing the same issue from happening again.

People won't remember every feature you shipped. They'll remember whether you stayed calm, communicated clearly, and helped when things were breaking.

Tools come and go. Understanding networking, Linux, debugging, and distributed systems will stay valuable throughout your career.

The best engineers explain problems clearly, document decisions, and keep everyone informed. Strong communication multiplies your technical ability.

Every extra dependency adds another way for things to fail. Choose the simplest solution that solves the problem well.

Great engineers aren't the ones with every answer. They're the ones who know how to investigate, ask good questions, and keep learning.

👋 Find me on Twitter | Linkedin | Connect 1:1

Thank you for supporting this newsletter.
Y’all are the best.