Send us your events. We'll tell you what breaks.
No dashboards to build, no alerts to configure, no query language to learn. You stream raw events; the platform learns what "normal" looks like for your business — including relationships between metrics you never wrote down — and tells you, in plain English, what changed and for whom.
Everything else — 6 services, 5 funnels — is running normally.
You send events. We do the rest.
There's nothing to configure. The moment your events start flowing, four things happen quietly and continuously — you never touch a knob.
You send events
Any meaningful thing that happens in your product — a signup, a checkout, a payment, a page view — with whatever attributes you already have (country, plan, platform…).
We learn your normal
For every metric, we learn what's normal by hour of day and day of week from your own history — and we discover the relationships between your events on our own.
We catch what changed
Spikes, silent drop-offs, error surges, slow weekly drifts — and the outlier segment behind them, even when the overall numbers look fine.
We explain it plainly
What changed, for whom, since when, and what to check — in a sentence a founder can read, not a metric a data team has to decode.
Why "zero configuration" is the whole point
Traditional monitoring asks you to know what to watch in advance — build the dashboard, set the threshold, write the alert. You can't set a threshold for a problem you didn't predict. bigdata learns your normal and watches everything, including the relationships you'd never think to define, so it catches the breakage you weren't looking for.
Six kinds of problems, in plain English
Every finding leads with impact and a next step — never a raw metric. Here's the range, with a real example of how each one reads.
A slice fails while totals look fine
The highest-value catch: a problem hiding inside a segment your averages smooth over.
People stop converting — no errors
A step in a funnel quietly stops converting. Nothing throws an error; users just leave.
Errors surge past normal
An error rate jumps well above what's normal for that metric — an active incident.
The 1%-a-week you never notice
A drift too gradual to feel day-to-day but that compounds into a bad quarter.
Things get slower
The slowest requests climb above their learned normal — a drag before it's an outage.
Traffic spikes or vanishes
A sudden surge, or a metric that normally has steady traffic going quiet — often an outage.
Built to answer one question: is anything wrong?
Your dashboard opens with an answer, not a wall of charts. Three tabs, and most days you never leave the first one.
Home — the answer
A plain-English verdict at the top (healthy, or the one thing that needs you), then everything grouped so you can skim it in 30 seconds: what's active, your funnels, slow trends, and traffic.
About 1 in 5 purchases (18%) from Brazil on Android are failing — vs near-zero for everyone else. It hides in your overall rate.
Signup → activation
onboarding · dropping
Activation rate, last 7 days
onboarding
Bars show where people drop; the line shows when it started slipping.
Funnels — your relationships, learned
The platform discovers your funnels on its own — cart→checkout, signup→activation — and shows each as a set of bars. A step turns amber the moment it drops below its normal. Flip any card to a conversion-over-time line to see when a slip started. You can also name any two events to watch yourself.
Findings (the second tab) is the full, filterable log of everything ever flagged — active, resolved, and trends — for when you want to dig.
How to read the graphs
Lines = a metric over time
Hover anywhere for the exact timestamped value. Two lines let you compare a failing segment against everyone else.
Bars = a funnel
Each bar is a step; its width is the % that made it there. A bar turns amber when that step drops below its learned normal.
Flip = where vs when
Funnel bars answer "where are we losing people." Flip to the line and you see "when did it start slipping" — a week flat, then a cliff.
Dashed line = normal
Where you see a dashed reference line, that's the learned normal for that metric. The solid line pulling away from it is the anomaly.
Live in about 10 minutes
Two keys, one HTTP call, and you're sending. No agent to install, no schema to define.
Get your keys
When your account is created you receive two things — keep them safe:
| Credential | What it's for |
|---|---|
| API key | Goes in your app — it's what sends us events. Treat it like a secret. |
| Dashboard token | Signs you into your dashboard, and only your dashboard. |
Send your first event
An event is any meaningful thing in your product. One POST from anywhere, or one line with the SDK:
# batch many events in the array; a 202 means we've got it curl -X POST https://ingest.ddataai.com/v1/events \ -H "Authorization: Bearer YOUR_API_KEY" \ -H "Content-Type: application/json" \ -d '{"events":[ {"service":"checkout","eventName":"purchase", "attributes":{"status":"ok","country":"us","platform":"ios","amount_cents":4200}} ]}'
var client = BigDataClient.builder() .endpoint("https://ingest.ddataai.com") .apiKey("YOUR_API_KEY").build(); client.emit("checkout", "purchase", Map.of( "status","ok", "country","us", "platform","ios", "amount_cents",4200)); // emit() never blocks and never throws — safe in your request path.
Label events well (optional, but worth it)
Attributes are free-form — send anything. Two keys unlock the most, and one rule of thumb makes segments work:
status = "ok"/"error" powers error detection · duration_ms = a number powers latency.
Useful dimensions: country platform plan version age_band. Send bands, not raw numbers — age_band:"18-25" is sliceable; age:23 isn't.
Open your dashboard, then sit back
Sign in at ingest.ddataai.com/login.html with your work account, or with your account name and dashboard token. You'll see live traffic within a minute or two. As events flow, the platform learns your normal (about an hour), discovers your funnels, and starts telling you what breaks — no configuration, ever.
Questions, answered
How long until I see something useful?
Live traffic shows immediately. Baselines (your "normal" per hour and day) settle in about an hour. Learned funnels and segment findings appear as your traffic builds — usually the same day for a product with real volume.
Do I have to tell it what to watch?
No. That's the point. You send events; it watches every metric and the relationships between them, and learns what's normal on its own. You can optionally ask it to watch a specific pair of events (in the Funnels tab), but you never have to.
What counts as an "event"? What should I send?
Anything meaningful: signups, logins, checkouts, payments, page views, searches. Send the events that matter to your business, with the attributes you already have. More context (country, plan, platform) means richer findings — but even bare events work.
How much traffic does it need?
It shines on a product with real transactions — think thousands of events a day across the metrics you care about. Very low-traffic or pre-launch apps have less for it to learn from, so findings are sparser; that's honest math, not a limitation we can configure away.
Is my data private?
Yes. Your data is scoped to your account — your token only ever opens your own tenant. We read the events you send us to learn your normal and surface findings; nothing is shared across customers.
Will it drown me in alerts?
No. It leads with the one or two things that actually need you and groups the rest. It suppresses repeats of an ongoing incident and caps how much it'll surface in a day — a quiet dashboard is the normal state.
What doesn't it do?
It's not an APM tracer or a log search tool — it won't show you a stack trace. It tells you what changed in your business metrics and for whom, so you know where to look. Pair it with your error tracker for the line-of-code detail.
Send one event. See what we find.
Setup is about ten minutes and you don't configure anything after. When something matters, you'll know — with the why, not just a red number.
Get startedQuestions? Reply to your onboarding email — a real person answers.
bigdata · zero-setup event intelligence