CEDX Issues · Projects & Delivery
Issues tracks the work from intake to done with the clocks on: SLA attainment 88%, 25 breached clocks named row by row, cycle outliers flagged at z=4.44, and a triage queue with a named human on duty.
The labels catalog has a policy, not just tags: p0-watch is “Leadership-watched P0 stream” with 40 open; unused labels read “candidates to archive.” There are zero. Somebody prunes.
Runs on demo data. The SLA table still prints raw floats on some rows — see Limits; we would rather show it than crop it.
What it is
Nine items wait in the intake queue; a named person is on duty — Ravi, week 32, incident-first rotation. Each row arrives with a suggested component, owner and confidence, and duplicates are flagged at title similarity ≥ 30% before they enter the queue.
Four workflows — Standard delivery, Bug lifecycle, Incident path, Support SLA — with states and named transitions, not a free-for-all: Start, Submit review, Block, Unblock, Accept, Request changes, Quick close. 198 issues ride the default; the Support SLA workflow sits at zero and shows it.
SLA attainment reads 88% with the 25 breached clocks listed individually — issue key and hours over. Cycle outliers get a z-score: CHK-2401 at z=4.44, eight days against a two-day baseline, P0 bug, in progress.
Product tour
Not a mockup and not a concept deck. This is what opens at /app/issues.
01 — Overview
Queue health reads 78% and shows the components: SLA attainment 88%, 25 breached clocks, 148 nodes at high dependency risk, 169 cycle outliers. The alerts are root-caused — symptom, cause, fix — and the worst row is named twice: CHK-2401, dep risk 98 and a cycle outlier at z=4.44.
02 — Triage
The intake queue suggests routing from component ownership — checkout, core platform, inventory, data plane, mobile app — and prints its confidence. The person on duty accepts, re-routes or defers; nothing auto-commits to the backlog.
03 — Labels
Twenty-four labels, each with a description: p0-watch is the leadership-watched stream, blocked means external or internal wait, compliance covers audit, retention and policy. Open and total counts sit on every row — and the unused-label count is zero.
04 — Workflows
Standard delivery runs Story, Task and Spike through Open, In Progress, In Review, Blocked and Done — 32, 42, 40, 13 and 71 issues respectively. The transitions are named for the human action: Block, Unblock, Request changes, Quick close.
Who runs it
Roles, not references. We have no named customers yet, so nobody in these photographs is quoted, credited or claimed as one.
Works the intake queue on rotation: nine rows, each with a suggested route and a confidence, and a decision to accept, route or defer before the backlog sees it.
in queue · 9
Owns the labels catalog and the workflows — 24 labels with descriptions, zero unused, and transitions named so a new engineer reads them without a walkthrough.
labels · 24 catalogued
Watches the clocks: 25 breached SLA rows and the z=4.44 outlier are the stand-up agenda, not a monthly report.
SLA attainment · 88%The shape of it
Every figure below is legible in the captures above. Nothing here is a projection of your estate — it is the state of the demo data.
How it runs
A report lands in the triage queue with a suggested component, owner and confidence. A named human on duty accepts, routes or defers — duplicates are flagged first.
The issue joins the catalog properly tagged: p0-watch, blocked, compliance — 24 labels with descriptions, so the tag means the same thing to everyone.
It moves through a named workflow — Open → In Progress → In Review → Done — with Block and Unblock as first-class transitions, not comments.
SLA attainment, cycle time and dependency risk are scored continuously. A breach is a row in a table with hours over, not a surprise in a quarterly review.
One record
The same key appears on the triage queue, the sprint risk table and the delivery board. There is no export to keep in sync — the estate reads one record.
A committed issue is scored in the sprint's risk queue with its top signal — the same row, read for planning.
CEDX ProjectsThe board's 340 issues and 32 blocked are this queue's rows, grouped for delivery rather than for intake.
CEDX DeskThe support reports in the intake queue — double-charges, login loops — arrive from Desk conversations, with the customer thread attached.
CEDX TestA fix that ships is a change that gets verified; the issue's lifecycle continues into the test run, not into a copy of itself.
Limits
Finding this out on the third call is worse for you than reading it here, and worse for us.
Start
Open it right now
Your queue, your workflows
Issues with the rest of it
Questions
Yes. Every screenshot is a capture of the running build and you can open the same build at /app/issues. It runs on demo data.
From component ownership. Each intake row suggests a component and owner with a confidence score — 94% on the top row, 61% near the bottom — and the person on duty decides. Duplicates are flagged by title similarity, with the percentage shown.
An issue whose cycle time is far off its baseline, scored as a z-value. CHK-2401 reads z=4.44 — eight days against a two-day baseline for a P0 bug — and the alert names the issue, the score and the baseline.
Where an issue sits in the blocking graph. CHK-2401 scores 98 because it blocks one issue at depth 5 while in progress; 148 nodes in the workspace sit at high risk by the same measure.
Yes — the demo workspace runs four: Standard delivery (Story, Task, Spike), Bug lifecycle, Incident path and Support SLA. Each has its own states and named transitions, and the issue counts per workflow are on the screen.
No certification has been issued. What we can evidence about hosting, encryption, tenant isolation and production access is written up on the security page.
Live build, demo data, no card. Then check what your oldest open issue would score on the outlier table.