Can I vibe code Segment?

Build a high-throughput event ingestion proxy and data pipeline using Node.js Fastify, Upstash Redis, and ClickHouse.

1. API Specification:
Implement a Segment-compatible HTTP REST API accepting payloads at `/v1/track`, `/v1/identify`, `/v1/page`, and `/v1/batch`. Define standard JSON schemas enforcing `anonymousId`, `userId`, `event`, `properties`, `context`, and `timestamp` fields.

2. Ingestion Pipeline:
Upon receiving an HTTP event, validate the payload against the schema. Attach an ingestion timestamp and push the record to a Redis Stream queue (`events:stream`). Immediately respond to the client with `202 Accepted` in under 15ms. Never perform blocking HTTP calls to downstream services inside the primary request cycle.

3. Storage & Worker Engine:
Implement a background worker thread that consumes from `events:stream` in micro-batches (500 events or every 1 second). Bulk-insert raw JSON events into a ClickHouse table optimized for time-series analytics (`events_raw`).

4. Destination Dispatcher:
Create an extensible plugin structure for destination routing. Build connectors for (a) Google Analytics 4 Measurement Protocol and (b) Meta Conversions API. Dispatchers must consume from the queue, convert Segment event schemas to destination formats, and execute outbound HTTP requests.

5. Retries and Dead Letter Queue:
If a destination HTTP request fails with 429 or 5xx status codes, implement exponential backoff retries (up to 5 attempts). If retries are exhausted, write the payload to a ClickHouse Dead Letter Queue (`events_dlq`) with error context. Create an admin endpoint `/v1/admin/retry-dlq` to re-drive failed events.

6. Out of Scope:
Do not build a graphical configuration dashboard, complex visual drag-and-drop workflow builders, or multi-tenant user authentication.

$ each button prefixes agent-specific run instructions · build your own product, never copy proprietary code, trademarks or designs