Analytics & Data

Relay provides built-in analytics tools to measure user behavior, track custom business metrics, and export data to external systems.

Funnel Analysis

Define multi-step event funnels to measure conversion rates.

Creating a Funnel

Go to your app → More → Observe → Funnels → Create Funnel.

A funnel is a sequence of steps, each defined by a channel pattern and event name:

{
  "name": "Checkout Flow",
  "steps": [
    {"channel": "page.*", "event": "view"},
    {"channel": "cart.*", "event": "add"},
    {"channel": "checkout.*", "event": "start"},
    {"channel": "checkout.*", "event": "complete"}
  ],
  "window_minutes": 60
}

How It Works

1. Relay tracks which socket_id values appear at each step
2. Only users who completed step N are checked for step N+1
3. The window limits how long between first and last step
4. Results show counts and conversion rates per step

Reading Results

Conversion is always relative to step 1 (not the previous step).

Custom Metrics

Define your own metrics from event payloads.

Creating a Metric

Go to your app → More → Observe → Custom Metrics → Create Metric.

Example Metrics

Daily Revenue:

• Aggregation: sum
• Channel: orders-*
• Event: order.completed
• Field: data.amount

Average Response Time:

• Aggregation: avg
• Channel: api.*
• Event: response
• Field: data.duration_ms

Error Count:

• Aggregation: count
• Channel: errors-*
• Event: *
• Field: *

Viewing Metrics

Each metric shows:

• Latest value — the most recent computed value
• Time series — 24-hour chart of values over time
• Sample count — how many events contributed

Metrics are computed on a rolling basis and stored as time-bucketed values.

Data Export Pipelines

Stream event data to external systems for warehousing, analytics, or compliance archival.

Supported Destinations

Creating a Pipeline

Go to your app → More → Observe → Data Exports → Create Pipeline.

Configure:

Destination Configuration

Each destination requires specific credentials:

S3:

Bucket: my-relay-events
Region: us-east-1
Access Key: AKIA...
Secret Key: ...
Prefix: relay/events/

BigQuery:

Project ID: my-project
Dataset: relay_events
Credentials JSON: {...}

Kafka:

Brokers: kafka1:9092,kafka2:9092
Topic: relay-events
SASL Username: (optional)
SASL Password: (optional)

Webhook:

URL: https://your-server.com/ingest
Secret: (auto-generated for HMAC signing)

Pipeline Status

Each pipeline shows:

• Events exported — total count
• Last exported — timestamp of last successful export
• Enabled/Disabled — toggle without deleting
• Flush interval — how often batches are sent

Data Format

Exported events follow this structure:

{
  "id": "01HQ...",
  "channel": "orders-us",
  "event": "order.created",
  "data": "{\"amount\": 99.99}",
  "payload_size": 24,
  "fired_at": "2024-03-15T14:30:00Z",
  "app_id": "01HQ..."
}

API Access

All analytics data is also available via the Intelligence API:

# Query events
POST /api/v1/apps/{appId}/intelligence/query
{"query": "SELECT COUNT(*) FROM events WHERE channel LIKE 'orders-*' SINCE 24h AGO"}

# Get cost attribution
GET /api/v1/apps/{appId}/intelligence/cost-attribution

# Get latency percentiles
GET /api/v1/apps/{appId}/intelligence/latency?channel=orders

See the Intelligence & Analytics docs for the full API reference.