CEDX Integrations · Developer & low-code
Integrations watches the connections themselves: 48 connectors, 72 webhook endpoints and 56 OAuth apps, with failure storms clustered, token expiry counted in days, and every alert arriving root-caused.
The alert panel does not say “something failed”. It says: checkout.paid storm · 42% fail — origin 503 for about 40 minutes, retries amplified to 4.2×. That is the difference between monitoring and knowing.
Runs on demo data — fleet health 91 is the demo seed's, and says so on screen.
What it is
The Connectors table carries provider, category, auth type, runs per day, fail % and a 14-day sparkline per row. commerce-orders runs 42K a day at 0.4% fail; partner-webhooks degrades at 12.4%; two support connectors sit in auth fail at 16.7% and 18.2%. The fleet is sortable by the thing that is actually wrong.
56 OAuth apps with grant type, days to expiry, dependents, refresh failures and a risk score. billing-live sits at risk 99 in refresh-fail; crm-prod-oauth expires in 3 days with deal-won flows already intermittent. Expiry is a countdown you can see, not an incident you discover.
72 endpoints with deliveries per day, fail %, p99, retry multiplier and storm score. wh-checkout-paid delivers 184K a day at 42.1% fail — storm 99, retrying at 4.2×. A storm score of 99 is a clustered failure window, not a noisy endpoint.
Product tour
Not a mockup and not a concept deck. This is what opens at /app/integrations.
01 — Overview
Deliveries at 1.84M a day, run success 96.4%, 41 of 48 connectors active — and then the risk cards: 9 auth at risk, 12 open storms, 18 failed runs needing retry. The alert panel root-causes: an origin 503 for ~40 minutes amplified webhook retries to 4.2× and blocked 14 downstream deliveries.
02 — Connectors
48 connectors across commerce, payments, CRM, support, inventory, marketing, identity and HR. The top source by volume is commerce-orders; the top problem is crm-pipeline, in auth fail at 18.2%. Both facts are on the same screen.
03 — OAuth apps
56 apps with days, dependents and risk. The refresh-fail chip isolates 15 apps needing re-auth; the expiring chip finds crm-prod-oauth at 3 days. wms-cli-29 shows what good looks like: rotated, risk 58, 20 dependents intact.
04 — Webhooks
Endpoints sort by storm score: wh-checkout-paid at 99, wh-deal-won-70 at 92. The degraded band starts at wh-invoice-paid-14, 24.9% fail with a 762ms p99. Delivery volume, failure and latency live on one row per endpoint.
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 map: which systems connect, through which auth, and which connector is one expired token away from a bad Monday.
48 connectors · 544K runs/day
Watches deliveries and storms: 12 open failure windows, 18 failed runs to retry, and the retry multiplier that turns an origin 503 into a flood.
72 endpoints · 21 stormy
Rehearses the rotation before it is forced: a cloned environment is where a new token or a re-mapped webhook gets proven first.
56 OAuth apps · 30 at riskThe 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 connector registers with its auth — OAuth2, API key, HMAC, Basic or JWT — and starts accruing runs, failures and a 14-day sparkline.
The fleet view rolls deliveries, success and storms into one health score: 91 today, with 12 clustered failure windows open and named.
OAuth apps carry days-to-expiry and refresh-fail counts, so crm-prod-oauth at 3 days is a task this week — not an outage next week.
Failed runs queue for retry: 18 in the last 24 hours, each marked need-retry rather than dropped without a trace.
One record
Integrations is not a third-party broker stapled on. The events, tables and runtimes it feeds are the estate's own.
Inbound connections resolve to routes and keys — the same registry where blast radius and quota risk are scored.
CEDX TablesSynced records land in bases with owners, sensitivity labels and health scores, not in an unmapped staging area.
CEDX ScriptsSchedules listen for the events connectors deliver — order.settled and invoice.paid are what automation fires on.
CEDX SandboxA cloned environment is where a re-mapped connector proves itself before production traffic moves.
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 first connectors
Integrations 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/integrations. It runs on demo data, and the header says “demo seed” on screen.
A clustered failure window, scored per webhook endpoint. When an origin 503'd for about 40 minutes, retries amplified to 4.2× and 14 downstream deliveries blocked — the storm score of 99 on wh-checkout-paid is that event, not a noisy-week average.
The OAuth apps table counts days to expiry and refresh failures per app, with dependents listed — crm-prod-oauth shows 3 days, 18 dependents, risk 81. The at-risk chip collects everything 14 days out or already failing.
The demo workspace uses generic provider labels, and this page deliberately does not publish a provider catalog. That conversation is part of a pilot scoping call, where we can speak to your specific systems.
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 count how many expiring tokens your current setup is tracking.