1
The pipeThrough() method of the ReadableStream interface provides a chainable way of piping the current stream through a transform stream or any other writable/readable pair.
2
Piping a stream will generally lock it for the duration of the pipe, preventing other readers from locking it.

javascript

Difficulty: 5/10

Topics: web-streams-api, transform-streams, backpressure

0-2 years experience

1
You're fetching a large JSON file from an API and need to parse it line-by-line without loading the whole thing into memory. How would you use pipeThrough with a transform stream to do this?
2
A teammate wrote code that does fetch(url).then(r => r.body.pipeThrough(decompress)).then(process) but it's not working. What's missing and how would you fix it?

2-5 years experience

1
You're building a log ingestion pipeline: HTTP request → decompress → parse JSON → validate schema → write to DB. The validation step occasionally throws on malformed records. How do you structure the pipeThrough chain so bad records are logged and skipped without stopping the whole pipeline?
2
Your team migrated from Node's legacy stream.pipeline() to Web Streams pipeThrough. After deployment, memory usage spiked on large file uploads. What's a likely cause and how would you debug it?

5-8 years experience

1
Design a reusable TransformStream that implements rate-limiting for any pipeThrough pipeline. It should respect backpressure, allow burst tolerance, and expose metrics for observability. Walk me through the key methods you'd implement.
2
You're streaming video transcoding output through multiple pipeThrough stages (decrypt → decode → filter → encode → encrypt). The pipeline stalls intermittently under load. How do you instrument and diagnose whether it's a backpressure deadlock, a slow transform, or buffer bloat?

8+ years experience

1
Your org has 50+ services using a mix of legacy Node streams, Web Streams, and async iterators. You're tasked with defining a cross-team streaming standard. What criteria would you use to decide when pipeThrough is the right abstraction vs. when to allow alternatives? How do you handle gradual migration?
2
A critical data pipeline uses pipeThrough for exactly-once processing semantics across service boundaries. The transform streams are stateful and need to survive restarts. How would you architect checkpointing and recovery without breaking the composable pipeThrough model?

❓How does error propagation work when the transform stream throws?
❓What happens to backpressure signals when you chain multiple pipeThrough calls?
❓When would you choose pipeThrough over manually handling readable/writable events?