Event spine
Server-side collection that proves it is working.
Instrumentation is usually the step that stalls a data project. Quantiv keeps it short: a few lines of Python, a browser helper that installs itself, and a health panel that shows each environment reporting in.
Instrumentation health
Norvelle
| Env | Events | Last heartbeat | Helper |
|---|---|---|---|
| prod | 1.84M | 14s ago | v1.4.2 |
| stage | 42.1K | 6h ago | v1.3.9 |
| test | 8.9K | 52s ago | v1.4.2 |
Illustrative data
Install
Three calls cover almost everything.
track records an event against your taxonomy. identify attaches traits to a person. alias joins an anonymous visitor to the account they later sign in to.
Each call is wrapped in a versioned context envelope with page, campaign, click identifiers, and consent signals, so attribution and forwarding have what they need without extra work. The SDK queues and batches in the background and fails open: if anything goes wrong, your application carries on.
import quantiv_sdk
# Reads QUANTIV_INGESTION_KEY / QUANTIV_INGESTION_URL from the environment
quantiv_sdk.init()
def complete_order(order):
quantiv_sdk.identify(order.customer_id, {"plan": order.plan_code})
quantiv_sdk.track("purchase", {
"order_id": order.id,
"order_value": float(order.total),
"currency": order.currency,
"email": order.email,
})
# Calls never raise into your app and never block the request.Click context
The browser helper installs itself.
Server-side events are reliable, but the click that started a visit happens in the browser. The Quantiv helper captures the page, referrer, UTM parameters, gclid, and fbclid, and hands them to the server-side envelope.
On Django, Flask, FastAPI, and Starlette the SDK injects the helper into rendered pages when it starts. Single-page and static frontends add a one-line snippet instead.
- Captures UTMs, gclid, fbclid, and referrer on landing
- Sends a heartbeat so its presence is verifiable
- Checks its origin against the base URL you registered
Instrumentation health
Norvelle
| Env | Events | Last heartbeat | Helper |
|---|---|---|---|
| prod | 1.84M | 14s ago | v1.4.2 |
| stage | 42.1K | 6h ago | v1.3.9 |
| test | 8.9K | 52s ago | v1.4.2 |
Illustrative data
Environments
Test traffic never reaches production numbers.
Every app has test, stage, and prod environments. Each has its own write keys, its own validation rules, and its own retention window.
Separate keys
A key identifies both the app and the environment, so an event can only land where it was meant to.
Stricter in prod
Unknown events are rejected in prod and flagged in test and stage, so mistakes surface while they are still cheap.
Retention per environment
Test and stage keep short, fixed windows. Production retention is a setting your organisation controls.
Under the hood
Built to keep up with volume.
Events are not stored as application records. They move through a pipeline designed for append-heavy workloads.
- 01
Authenticated ingestion
Each request is checked against its write key and per-key rate limits.
- 02
Validation
The payload is checked against the Active definition for that event before anything is accepted.
- 03
Queue and batch write
Accepted events are queued and written in batches to an append-only event store.
- 04
Usage limits without silent loss
If traffic exceeds your plan, it is throttled and your administrators are alerted. Nothing is silently dropped or silently billed.
Questions
Common questions.
Which languages are supported?
The server SDK is Python. The browser helper works with any frontend, either injected automatically by the SDK or added as a snippet.
Do we still need the Meta pixel?
You can keep it. Quantiv deduplicates server-side conversions against pixel events, so running both improves coverage without double counting.
How do we know instrumentation is complete?
The Instrumentation Health panel shows the last heartbeat, helper version, and origin match per environment. The instrumentation agent also reports events arriving that your taxonomy does not yet define.
See it on your own numbers.
A two-week pilot runs your conversions through Quantiv alongside what you report today. The difference between the two is the business case.