Formula
PDRP = (PRs with ≥1 failing Lens under this pillar
÷ total PRs evaluated)
× 100Thresholds
Healthy
Caution
Critical
Reading it
PCR measures the pass rate of individual Lens evaluations. PDRP measures how broadly a pillar's failures are distributed across PRs. A high PDRP means many different contributors are diverging from this pillar's intent — not just one team failing repeatedly.
This distinction drives the remediation. If PDRP is high but failures are concentrated in a small set of teams, the problem is a one-team enablement gap. If PDRP is high and failures are spread across many teams and contributors, the problem is systemic — either a training gap, a tooling gap, or Lenses that no longer reflect the org's actual architectural intent.
A VP Eng asking "is this one team or everyone?" is asking the PDRP question. An architect asking "before I recalibrate these Lenses, how widely is this actually failing?" is also asking the PDRP question. PCR and PDRP together give both the rate and the reach of a pillar's conformance problems.
How rkito produces it
- 01
Each PR is evaluated against applicable Lenses
The CDA pipeline runs every Lens associated with the target pillar against the PR diff
- 02
PR is flagged if any Lens under this pillar fails
A single failing Lens is sufficient — the PR counts toward the PDRP numerator regardless of how many other Lenses passed
- 03
Count PRs with ≥1 failure under this pillar
The numerator grows once per PR, not once per failing Lens evaluation — this is the key difference from PCR
- 04
Divide by total PRs evaluated, multiply by 100
Computed per pillar over a rolling 30-day window, producing a 0–100 rate where lower is better
- 05
Navigate for detail
Go to /changecontrols/designaudits, filter by FAIL, identify which teams are submitting the failing PRs
Who reads it
PDRP answers the first triage question after a pillar's PCR drops: is this concentrated in one team, or distributed across the org? Concentrated PDRP with low breadth points to an enablement problem with a known owner. Distributed PDRP points to systemic drift that requires a different response.
A low PCR with a high PDRP means failures are spread widely across contributors — which can indicate Lenses firing on benign patterns before it indicates teams failing at architecture. PDRP gives the architect the evidence to confirm whether a Lens recalibration is warranted before initiating a CHI.
How it misleads
What to do at each threshold
Step-by-step response playbook for each signal state.
Pillar failures are contained. Fewer than one in five PRs is touching this pillar's intent. No immediate action required.
- Monitor on a monthly cadence. Confirm the trend is stable or improving before standing down.
Failures are spreading. Identify whether this is a concentrated or distributed problem before choosing a response.
- Navigate to
/changecontrols/designaudits, filter by FAIL, identify which teams are submitting the failing PRs. - Failures concentrated in one team: run a targeted enablement session with that team and the pillar owner.
- Failures distributed across many teams: initiate a Lens recalibration review with the pillar owner before assuming the problem is behavioural.
More than four in ten PRs are triggering failures under this pillar. Convene the pillar owner immediately — but diagnose before escalating.
- Open each failing Lens at
/steerings/lenses— determine if the prompt still accurately reflects current architectural intent. - Miscalibrated Lenses get updated via CHI. Do not escalate to engineering leadership until miscalibration has been ruled out.
- Genuine drift confirmed: run an awareness campaign across contributing teams and escalate to engineering leadership with the audit evidence.