Engineering intelligence

Every board has an Intelligence tab — a configurable, ClickHouse-backed console of widgets covering delivery, flow, quality, and AI adoption. Pick a period, add the widgets you care about, and drill through to PRs and developer profiles.

deckgauge · Board · Intelligence
Lead time (p50)
31h
High
Deploy freq
4.2/wk
Elite
Change-fail
9%
High
Restore time
3.5h
Medium
PR velocity (weekly merged)
AI-assisted PRs
46%
last 30 days
Ticket coverage
88%
PRs linked to a ticket

A selection of widgets — DORA tiles, PR velocity, and AI adoption.

Powered by ClickHouseRaw events (PRs, reviews, commits, work items, transitions) are streamed into ClickHouse, where materialized views pre-aggregate the metrics — so dashboards stay fast even across large histories. It all runs in your own stack.

Widget catalog

Widgets are grouped by theme. Each declares which sources it needs (Jira / GitHub / GitLab / ADO); widgets without their required source are greyed out.

Delivery & DORA

  • DORA Metrics — lead time, deploy frequency, change-failure rate, time-to-restore, with Elite/High/Medium/Low tiers.
  • Lead Time for Changes — first-commit → merge (p50).
  • Delivery Trend — annotated daily delivery with calendar bands and peak highlighting.

Speed

  • PR Velocity (weekly merged) · PR Cycle Time Trend (weekly p50) · PR Cycle Time Scatter (one dot per PR, click through).
  • Review Pickup Time (opened → first review) · PR Size Distribution (XS–XL).
  • Velocity Leaderboard · Merge Frequency per Developer · Commits per Developer (lines changed + AI-assist %).

Flow

  • Completion Rate · Completion Trend · Issues Opened vs Closed.
  • Work in Progress (WIP) · Throughput & Cycle Time.

Planning

  • Iteration Planning Accuracy — % of planned work finished in scope.
  • Velocity with Confidence Band (±1σ) · Initiative Risk Radar (on-track / at-risk / overdue) · Backlog Age.
  • Investment Allocation — delivered effort split across feature / bug / tech-debt / KTLO.

Quality

  • Rework Rate — % of commits re-touching lines within 21 days.
  • Bug Rate · Review Quality Index (coverage, dwell, instant-merge, comments, ticket links) and its trend.

AI adoption

  • AI-Assisted PR % · AI Adoption (monthly AI PR% / commit%) · Review Mix (Bot vs Human) · Reviewer Participation.

Board health & correlation

  • Status Distribution, Status by Group, Items by Owner, Stuck in Progress, Blocked / Stale Items.
  • Ticket Coverage Rate — % of merged PRs linked to a ticket.

Comparison view

A multi-board Comparison puts boards side by side as columns — Compare: Flow, Delivery, and Review Quality. Metrics that aren't comparable across tools are greyed out rather than shown misleadingly.

Honest metricsDeploy frequency, change-failure rate, and restore time are proxies until you connect a deployment/incident source — Deckgauge labels them as such rather than overstating precision.