Skip to content

Connect a source

A source is one of your other systems — the place your customer data already lives. Connecting a source lets Pulse pull that data in automatically, so it can be united into customer profiles and used to build audiences. You can connect a SaaS tool, a streaming feed, an HTTP API, or simply upload a file.

This guide walks you through connecting your first source and getting data flowing.

Before you start

Your data lands in a dataset. It helps to have a schema ready so Pulse knows the shape of your data — see Create a schema — but you can also let Pulse land your records and describe them afterward.

Open the Sources screen

  1. Sign in to Pulse at pulse.azothedge.com.
  2. In the left navigation, under Ingestion, select Sources.
  3. You'll see the catalog of connectors you can choose from.
The Sources catalog in Pulse showing available connectors
The source catalog: pick the kind of system you're bringing data in from.

Choose a connector

Pick the connector that matches where your data lives:

  • Stripe — bring in customers and charges from your payments platform.
  • Apache Kafka — connect a streaming feed of real-time events.
  • HTTP API — pull records from a web service that Pulse can call.
  • Local file — upload a file (CSV, JSON, or similar) straight from your computer.

Select the connector you want to start with.

Setting up an HTTP API source connection in Pulse
Each connector asks only for what it needs to reach your system.

Create and test a connection

A connection is the saved link to your system — its name and the credentials Pulse uses to reach it.

  1. Give the connection a clear Name — for example, Production Stripe.
  2. Enter the details that connector needs (for example, an access key, an endpoint, or sign-in credentials). Pulse stores these securely.
  3. Select Test to confirm Pulse can reach your system. Fix any details if the test reports a problem.
Configuring an Apache Kafka streaming connection in Pulse
Test the connection before you go further, so you know it's reachable.

Keep credentials current

If a credential changes in the other system, update it on the connection here and test again. Everything that relies on the connection will use the new details from then on.

Create a dataflow

A dataflow describes what to pull and where it should land.

  1. With your connection selected, create a new dataflow.
  2. Choose the kind of record to bring in — for example, customers or charges.
  3. Pick the target dataset the records should land in, or create one for them.
  4. Choose a schedule for how often Pulse pulls the data:
    • Manual — only when you run it yourself.
    • Hourly — automatically, once an hour.
    • Daily — automatically, once a day.

Preview before you commit

Select Sample to pull a few records and preview them. It's the quickest way to confirm you're bringing in the right data, mapped the way you expect.

Run it and check the results

  1. Select Run to pull data right away.
  2. Open the dataflow's run history to see each run — when it ran, how many records were fetched, and how many were accepted.
  3. If records were rejected, the history points you to the reason so you can adjust and run again.

From here your data lands in its dataset, and — once that dataset is enabled for unified profile — starts building customer profiles.

Next steps

  • Manage your datasets

    See where your records landed, preview them, and enable them to build profiles.

    Manage datasets

  • Describe your data

    Use a schema to tell Pulse the shape of the data you're bringing in.

    Create a schema