CEDX Monitoring · IT & Security
CEDX Monitoring watches 48 services, 320 probes and every error budget — and instead of paging you with a symptom, it hands you the cause, the blast, and the next action, scored.
Checkout p99 is burning 14 times its budget rate. The alert says why — a dual-write path lagging since Monday's deploy — and the next action is scored 94: cut dual-write traffic to 20%.
Runs on demo data — Northline Production is the software's sample tenant, not a customer.
What it is
The service catalog carries tier, status, RPS, p99 and error rate per row — ledger-api at 7.9K requests a second with p99 119 and 3.1% errors is critical; the 10 services with no SLO are listed as a coverage gap, not quietly unwatched.
Checks are typed — HTTP, TCP, DNS, SSL, browser — with latency, uptime and failures per probe. The checkout browser probe takes 1,504ms at 97.3% uptime; the edge TLS expiry probe is up at 100.0%.
320 alert conditions, 39 firing, 64 with a noise score over 80. The table shows fires, pages and noise per alert — an idempotency alert at 116 fires and 19 pages with noise 99 is a mute conversation with evidence.
Product tour
Not a mockup and not a concept deck. This is what opens at /app/monitoring.
01 — Overview
17 open incidents, 39 alerts firing, 20 checks down, 16 SLOs with under 40% of budget left. The root-caused panel does the handover: payments SEV-2 still mitigating — a missing circuit breaker on one card network — and the partner 429 storm traced to a batch job without jitter.
02 — Services
Sort by error rate and the argument ends: which service burns fastest this week is a sortable column, and the 6 critical-status services sit at the top with their p99 and 7-day trend.
03 — Checks
Search checks by service or target; the critical-path card counts 14 probes down or degraded. Every probe shows what it is — SSL expiry, DNS, a full browser run — with its own latency and uptime.
04 — Alerts
The ask bar suggests the questions: which alerts page the most, which noisy alerts we can mute, which criticals saw no pages this week. The table answers them with fires, pages and noise score per row.
Who runs it
Roles, not references. We have no named customers yet, so nobody in these photographs is quoted, credited or claimed as one.
Gets paged with the cause attached, works the scored next actions, and owns the incidents the agent has already investigated.
open incidents · 17
Owns a row in the catalog: the SLO, the budget burn, the probes on the critical path — and the 10-service no-SLO gap if their service is in it.
no SLO · 10
Prunes the 64 noisy alerts, reviews the 42 active mutes, and keeps mute debt from becoming blindness.
noise ≥80 · 64The 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
320 checks watch the paths that matter — SSL expiry to full browser runs — with latency and uptime per probe.
Signals fire against SLOs and error budgets; the golden-signals chart carries deploy markers so cause and effect share a timeline.
The agent works every open incident — 17 of 17 — and the alert lands root-caused: symptom, cause, and the deploy it started with.
Next actions arrive scored: cut dual-write traffic to 20% at 94, silence the cache-flap cluster at 91, ship the circuit breaker to close the SEV-2.
One record
A firing alert is the start of work, not the end of it — and the work lives in the same estate that watches the signal.
A SEV-1 becomes a major blast with a commander and a correlated incident set — the monitoring alert is the root cause it starts from.
CEDX NetworkThe partner 429 storm and the shadow-route findings are flows Network governs — the alert and the policy share a record.
CEDX EndpointA service's p99 regression after a deploy correlates with the devices and agents Endpoint watches — same estate, same timeline.
CEDX SIEMAuth and lateral-movement signals feed the SOC queue with verdicts attached — operational and security signals share one stream.
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 services, your SLOs
Monitoring 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/monitoring. It runs on demo data — Northline Production is the sample.
The alert names the symptom, the cause and the fix: checkout p99 burning 14× is traced to the dual-write path lagging since Monday's deploy, and the next action — cut dual-write traffic to 20% — is scored 94. You can argue with all three on the screen.
Noise is a score per alert, with fires and pages counted. 64 alerts sit above 80 and are named as mute candidates; 42 mutes are active, and mute debt is tracked on the health ring so quiet does not become blind.
Typed probes — HTTP, TCP, DNS, SSL and full browser runs — 320 of them, each with latency, uptime and failure counts. Fourteen sit on the critical path and get their own card.
No certification has been issued. What we can evidence about hosting, encryption, tenant isolation and retention is written up on the security page.
Live build, demo data, no card. Then ask which of your alerts would survive a noise score.