@fastify/etag

@fastify/etag

A plugin for Fastify that automatically generates HTTP ETags according to RFC2616-sec13.

The plugin can optionally send a 304 status code when an ETag matches the if-none-match header.

Install

npm i @fastify/etag

Compatibility

Plugin version	Fastify version
>=6.x	^5.x
>=4.x <6.x	^4.x
^3.x	^3.x
>=1.x <3.x	^2.x

Please note that if a Fastify version is out of support, then so are the corresponding versions of this plugin in the table above. See Fastify's LTS policy for more details.

Example

'use strict'

const Fastify = require('fastify')
const Etag = require('@fastify/etag')

const app = Fastify()
app.register(Etag)

app.get('/', async (req, reply) => {
  return { hello: 'world' }
})

app.get('/manual-etag', async (req, reply) => {
  // This will disable automatic ETag generation
  // It will still return a 304 if the ETag matches
  reply.header('etag', '"foobar"')
  return 'world'
})

app.listen(3000)

Plugin Options

• algorithm: all hashing algorithms the Node.js crypto module supports, and 'fnv1a'. Default: 'sha1'.

• weak: generates weak ETags by default. Default: false.

Automatic 304 status codes

By default, the plugin sends a 304 status code when the ETag is equal to the Etag specified by the if-none-match request header.

This is often the desired behavior, but can be disabled by setting replyWith304: false.

Acknowledgments

The fnv1a logic was forked from Travis Webb (@tjwebb)'s https://github.com/tjwebb/fnv-plus, the actual implementation is contributed to fnv-plus by @desudesutalk in https://github.com/tjwebb/fnv-plus/pull/9, and adapted by @SukkaW to support buffers.

Benchmarks

Generating an etag will always be slower than not generating an etag. The generation speed also depends on the payload size and type (buffer or string):

• For very small payloads (< 2 kb), use 'fnv1a'
• For buffers above 2 mb, use 'md5'
• In all other scenarios, use 'sha1' (default)
• YMMV, see this issue where other algorithms such as crc32 for small payloads and murmurhash3-wasm for big buffers have performed better than the mentioned recommendations
• Any etag generation results in at least 10% less op/s (up to 50% less op/s for huge payloads)

License

Licensed under MIT.