Production diagnostic
What an audit memo looks like
The memo a production diagnostic hands back. Three ranked causes for a system that is stable at low load and degrades under sustained traffic: an N+1 query on the largest accounts, a burstable RDS instance out of headroom, and a deploy pipeline whose health check proves a task is running but not that it can serve. Written from an intake packet, with no shell or production access assumed. The company is fictional and every finding invented.
Node.js · AWS ECS · Deploy pipeline
-
Findings, ranked
Three findings ordered by blast radius, not by how interesting the code is. Each one carries the observed symptom, the evidence, how to verify it, the fix, and the effort.
-
A go / no-go you can act on
Two remediation tracks with scope and price, so the next decision is a budget line, not another meeting.
-
What I deliberately left alone
A short "not worth prioritizing now" section. Knowing what not to spend money on is half of what you are paying a senior engineer for.
What an audit memo looks like
Fictional sample · 4 pages
The audit is a standalone engagement. Implementation and verification can be scoped separately.