Page on symptoms, not causes
If users aren't hurting, it can wait until morning. If they are, the alert should say how badly.
~/loveth-oballe/about
About
I started on a university help desk, resetting passwords for twelve thousand students. Most of the job was cleaning up after access nobody had planned for, and that's still the problem I find most interesting.
From there I moved into systems administration, then DevOps, then SRE. Each move came from the same habit: when something broke twice, I wanted to fix the system that let it break, not just the incident.
Today I work where reliability and security overlap. An expired certificate, an over-privileged role and a noisy pager are usually one missing guardrail. I'd rather build that guardrail once than hold the line by hand.
How I work
The opinions I bring to a new team. Happy to argue any of them in an interview.
Access should expire by default
Permanent admin is a liability. Elevation that ends on its own is one nobody has to remember to revoke.
The secure path is the easy path
If the safe way takes longer, people route around it. Guardrails belong in the pipeline, not in a wiki.
Write it down for the next on-call
Runbooks, postmortems and ADRs are how a team stops relearning the same lesson.
Outside the pager
I record walkthroughs of my projects for YouTube, because explaining a system out loud is the fastest way to find the part I don't understand yet.