Telemetry
Live clickstream analytics powered by Kafka to ClickHouse. Switch time windows and inspect recent events.
Clicks (24h)
—
Unique sessions (24h)
—
Latency P50
—
Latency P95
—
Latency guide: P50 is the typical ingestion time (median). P95 shows the slower tail (95% of events are this fast or faster).
24h
Live Event Inspector
latest interaction events (max 200)Time (PST)EventElementPageSession
No interaction events yet.
Pipeline Health
Last event (PST)—
Refresh cadenceEvery 5s
Window24h
Data engineering signals
- Event schema + versioning
- Partitioning by session_id
- Streaming ingest into ClickHouse
- Operational latency monitoring (P50 / P95)
Top Clicked Elements
24hNo click data yet.
Top Pages
by eventsNo page data yet.
Data Contract
Event schema and contract
A simple event model keeps the telemetry pipeline consistent, queryable, and extensible.
Sample Event Payload
JSON{
"event_id": "0b7d3a9e-8d6a-4d7e-9b31-4ec1f16f7b7e",
"ts": "2026-02-26T18:42:11.231Z",
"session_id": "s_9f6a4c...",
"event_name": "button_click",
"element_id": "resume_download",
"page": "/",
"props": { "source": "nav" },
"v": 1
}Event Contract
Schemaevent_idUUID
Idempotency key for dedupe / retries
tsDateTime64(3)
Event timestamp emitted by app
session_idString
Visitor session key and Kafka partition key
event_nameString
Event type (page_view, nav_click, etc.)
element_idString
UI element identifier
pageString
Route/path where event happened
propsJSON/String
Extensible metadata payload
vUInt8
Schema version