Compression Floor
Infrastructure principles from first principles. Every system has a minimal configuration where it still produces its characteristic behavior. Strip everything else.
Read the full essay →From production
SRE practiced as software engineering.
A retainer for teams without a full SRE function.
The work ships as code: runbooks, SLOs with enforced error budgets, on-call rotations sized for two events per shift. Your team operates the artifacts.
Symptoms of an unowned floor
-
Incidents eat weekends
Pages land at 2am and the fix lives in one person's head. Monday is recovery, not shipping.
-
Cloud costs drift
The bill grows every quarter and no one can say which service did it. Tagging was always next sprint.
-
Releases ship on trust
Rollback is a manual procedure no one has rehearsed. Every release waits on the one engineer who remembers how it breaks.
-
Toil eats your seniors
The architects you hired spend the week on alert noise and manual ops. The work they were hired for waits.
How It Works
Diagnostic
A 30-minute call maps your infrastructure risk surface. I name what's costing you time, money, and reliability.
Retainer
A fixed-scope monthly engagement. Health reviews, incident response protocols, cost audits, pipeline hardening, scoped to your needs.
Outcomes
Fewer incidents, lower costs, faster deploys. Your team gets back to building. SLA tracking proves the ROI.
From the Lab
Resilience Patterns
Practical patterns for building systems that withstand unexpected failures and adapt to changing conditions. Drawn from production, validated under load.
Read the full essay →About infrawei
I’ve spent years running infrastructure at scale: scaling systems, building on-call culture, and turning operational chaos into repeatable playbooks. infrawei productizes that knowledge into retained engagements for teams who need operational excellence without the full-time headcount.
Principles
- Playbooks over heroics
- Measure before you automate
- Reliability is a team sport
- Every system has a compression floor
Map the floor before you commit.
A 30-minute diagnostic. I walk your infrastructure and put the real reliability, cost, and deploy risk in writing, whether or not you retain me.