Lenss to audit rfcs
Audit an RFC against the goal it claims to meet.
System design and architecture reviews that start from the success criterion, not the diagram.
Works on: RFCs and design docs · Architecture decision records · Capacity and migration plans
How the audit runs
01
Pin the objective
The success criterion the design must meet, stated as a measurable target, separated from the solution it proposes.
02
Quote what it takes for granted
Each assumption the design rests on, tied to the exact sentence that carries it.
03
Find the operational limit
The capacity, latency or cost ceiling that binds first, and whether the design accounts for it.
04
Compute the breaking point
From the RFC's own figures, the value at which the design stops meeting its criterion.
Typical findings
Average vs tail
Caches, queues and retries that improve the mean while the SLO is a percentile.
Hidden ceilings
Throughput, connection or budget limits stated once and ignored in the plan.
Coupling
Behaviour the design needs from other services or teams, never agreed.
Worked example
Source draft
591 charsRFC: Redis cluster in front of the catalog database
Platform team · status: proposed
Adversarial review
pinned · quoted · computedPinned objective
p99 catalog read latency under 150 ms at 18,000 req/s.
Underneath: The success metric is a tail percentile; the proposal reasons about the average request.
Hidden assumption
A 90% hit rate moves the p99. It does not: p99 is set by the slowest 1% of reads, and with 10% of reads missing, that slowest 1% is all misses.
“We expect a 90% hit rate based on the 80/20 access pattern we see in logs.”
If it’s wrong: The cluster ships, average latency falls, and the SLO still fails at the same 380–450 ms tail.
Flip threshold
Cache hit rate at which p99 can fall under 150 ms
| Miss path: Postgres + cache lookup | 384–454 ms |
| Every miss vs the 150 ms SLO | over, always |
| p99 ≤ 150 ms needs misses ≤ | 1.0% of reads |
| p99 at the expected 90% hit rate | ≥ 384 ms |
At 90% the SLO fails. The decision turns on hit rate above 99%, not on the cluster.
Cheapest test: Replay one peak hour of read logs against an LRU of the planned size and measure the miss rate before provisioning.
Used most by Engineers
Audit your own draft now.
Paste up to 24,000 characters. You get the objective, every quoted assumption, the tensions, the flip threshold and a hardened rewrite.