A young engineer once asked an SRE veteran, “What advice do you have for someone like me to succeed in this field?” The veteran replied: “Don’t ignore alerts. Don’t deploy on Fridays. And avoid all single points of failure.”

It’s tough to define perfect reliability, but easy to spot what kills it. Building resilient systems takes effort; breaking them happens fast. When aiming to excel as an SRE, flip the script: focus on dodging the pitfalls that drag you down.

Here are a few pieces of very bad advice for Site Reliability Engineers.

  • Ignore alerts until they turn into outages.
  • Deploy changes without testing or canaries.
  • Assume “it works on my machine” means it’s production-ready.
  • Prioritize new features over fixing toil.
  • Never automate repetitive tasks; manual is fine.
  • Treat on-call shifts as someone else’s problem.
  • Skip post-mortems; just move on from incidents.
  • Scale vertically every time, ignore horizontal options.
  • Discount chaos engineering as unnecessary chaos.
  • View monitoring as a checkbox, not a lifeline.
  • Assume users won’t notice brief failures.
  • Optimize for 100% uptime without balancing costs.
  • Let tribal knowledge replace documentation.
  • Forecast capacity with gut feelings, not data.
  • Maximize efficiency so there’s no slack for surprises.
  • Be reactive: wait for problems to find you.
  • Associate tool popularity with actual usefulness.
  • Envy other teams’ setups without knowing their trade-offs.
  • Pursue certifications over hands-on experience.
  • Let ego drive decisions in incident response.
  • Assume all failures stem from bad code, not ops.
  • Believe past stability guarantees future uptime.
  • Use uncertainty in metrics as an excuse to skip improvements.
  • Judge other SREs by their worst incidents and yourself by your best.
  • Assume learning stops after your first big outage.
  • View patience in debugging as wasted time.
  • Use tools as a status symbol instead of solutions.
  • Only learn from your own outages, not others'.
  • Have no plan for your own burnout.

Based on the excellent Very Bad Advice.