CEDX Events · Platform
Eighty active streams moving 24.8K events a second in the demo estate, with the numbers stream tooling usually rounds away: p95 consumer lag at 4.2 seconds, schema compatibility at 94%, and 1,842 dead letters grouped into 118 open queues.
The Live screen prints its own caveat: deterministic demo series — not live push. A console that labels its own demo data is one you can trust with the real thing.
Runs on demo data — the header says demo seed, and the Live screen says the series is deterministic. Aether Production is the software's demo workspace.
What it is
The Live screen rolls at 112K events a second with the hottest streams ranked — commerce.cart at 6.1K, checkout at 4.2K — and a recent-samples tail that says demo tail, not a warehouse on the screen. The caveat is printed where you cannot miss it.
Volumes breaks 30 days of ingest into the domains that produced it: commerce at 9.2M events, identity at 4.1M, catalog at 2.8M — down to system at 680K. Bytes per day and peak eps per stream are on the same screen, so capacity talk starts from evidence.
Pipeline health reads 91 and counts its own deductions: 14 schema violations, 118 open DLQ groups, degraded 17 of 80 streams. The highest-lag table names commerce.refunds at 10.5 seconds — the stream that needs attention is not hidden in an average.
Product tour
Not a mockup and not a concept deck. This is what opens at /app/events.
01 — Overview
Ingest at 24.8K a second against 23.9K delivered, a 96.4% success rate, and the annotations on the chart are events a team would remember: schema enforce turned on, a mobile SDK version, a fan-out lag spike. The health ring counts violations and dead letters, not just uptime.
02 — Live
A rolling bar a second, hottest streams with sparklines, and a samples table where a Cart Updated event at 412 bytes sits next to a Payment Authorized at 510. The screen tells you itself: deterministic demo series — not live push; use Streams for partition lag.
03 — Volumes
Thirty days of ingest split by domain, a weekly capacity projection against baseline, and the stream table underneath with eps and peak per row. Commerce dominates at 9.2M events; the long tail is visible too.
Who runs it
Roles, not references. We have no named customers yet, so nobody in these photographs is quoted, credited or claimed as one.
Owns the lag board: 17 degraded streams, the 10.5-second refund stream at the top of the p95 list, and the 118 open dead-letter queues.
open DLQ · 118
Owns schema health — 94% compatibility, 14 violations — and the 24 streams exposed to agents, of which only 14 are governed. The gap is their queue.
schema violations · 14
Owns the alerts: lag, schema and DLQ arrive root-caused, so the 2am page names the stream, not the symptom.
p95 lag · 4.2sThe 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, which the product itself labels as seeded.
How it runs
A producer writes to a stream — one of 80 active, owned by a domain, with its schema registered.
Schema compatibility is enforced: 94% pass, and the 14 violations are a list with owners, not a log mystery.
Consumers read at their own pace; lag is measured as consumer age, p95 at 4.2 seconds across the estate.
What cannot be delivered lands in a dead letter queue — 1,842 events across 118 groups, each group a fixable work item.
One record
Events is how the other products hear about each other — the streams carry what the rest of the estate does.
Event triggers are a first-class trigger type in Workflow — these streams are what those automations fire on.
CEDX SyncMoving events between systems is Sync's job; the lag and delivery numbers here are how you know it is working.
CEDX Agents24 streams are exposed to agents as context-layer reads — 14 of them governed, a distinction this console enforces.
CEDX NotifyThe alert streams — lag, schema, DLQ — are exactly the signals that become deliveries in Notify.
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 streams, your schemas
Events 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/events. The demo data labels itself: the Live screen prints deterministic demo series — not live push where you can read it.
As consumer age, with the p95 across the estate on the Overview header — 4.2 seconds in the demo — and the worst streams named individually: commerce.refunds at 10.5 seconds leads the highest-lag table.
They land in dead letter queues, counted on the header (1,842) and grouped for work (118 open groups). The pipeline health ring deducts for them, so a growing DLQ is a visible score change, not a silent log.
Compatibility of the events against their registered schemas — 94% pass in the demo estate, with 14 violations listed. The chart annotations show when enforcement was switched on, because that is when the number starts meaning something.
Streams made readable to AI agents as context — 24 in the demo estate. The governed subset (14) is the one with access rules attached; the console counts both, because the difference matters.
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 ask what your own p95 consumer lag is right now — and who knows it.