Understanding & Measuring HTTP Timings with Node.js

Understanding and measuring HTTP timings helps us to discover performance bottlenecks in client to server or server to server communication. This article explains timings in an HTTP request and shows how to measure them in Node.js.

Before we jump into HTTP timings, let’s take a look at some basic concepts:

• IP (Internet Protocol): IP is a network-layer protocol, deals with network addressing and routing. IP is responsible for delivering packets from the source host to the destination host based on the packet headers across one or more IP networks. It also defines packet structures that encapsulate the data to be delivered.
• DNS (Domain Name Servers): DNS is a hierarchical decentralized naming system used to resolve human-readable hostnames like risingstack.com into machine-readable IP addresses.
• TCP (Transmission Control Protocol): The TCP standard defines how to establish and maintain a network conversation between applications to exchange data. TCP provides reliable, ordered, and error-checked delivery of a stream of octets between applications running on hosts communicating over an IP network. An HTTP client initiates a request by establishing a TCP connection.
• SSL/TLS (Transport Layer Security): TLS is a cryptographic protocol that provides communications security over a computer network. SSL (Secure Sockets Layer) is a deprecated predecessor to TLS. Both TLS and SSL use certificates to establish a secure connection. SSL certificates are not dependent on cryptographic protocols like TLS, a certificate contains a key pair: a public and a private key. These keys work together to establish an encrypted connection.

Now let’s take a look at the timeline of a usual HTTP Request:

Timings explained:

• DNS Lookup: Time spent performing the DNS lookup. DNS lookup resolves domain names to IP addresses. Every new domain requires a full round trip to do the DNS lookup. There is no DNS lookup when the destination is already an IP address.
• TCP Connection: Time it took to establish TCP connection between a source host and destination host. Connections must be properly established in a multi-step handshake process. TCP connection is managed by an operating system, if the underlying TCP connection cannot be established, the OS-wide TCP connection timeout will overrule the timeout config of our application.
• TLS handshake: Time spent completing a TLS handshake. During the handshake process endpoints exchange authentication and keys to establish or resume secure sessions. There is no TLS handshake with a not HTTPS request.
• Time to First Byte (TTFB): Time spent waiting for the initial response. This time captures the latency of a round trip to the server in addition to the time spent waiting for the server to process the request and deliver the response.
• Content Transfer: Time spent receiving the response data. The size of the response data and the available network bandwidth determinates its duration.

How do HTTP timings help to discover bottlenecks?

For example, if your DNS Lookup takes longer time than you expected, the issue might be with your DNS provider or with your DNS caching settings.

When you see longer Time to First Byte durations, you should check out the latency between the endpoints, but you should also check out the current load of the server.

Slow Content Transfer can be caused by inefficient response body like sending back too much data (unused JSON properties, etc.) or by a slow connection as well.

Measuring HTTP timings in Node.js

To measure HTTP timings in Node.js, we need to subscribe to a specific request, response and socket events. Here is a short code snippet how to do this in Node.js, this example focuses only to the timings:

const timings = {
    // use process.hrtime() as it's not a subject of clock drift
    startAt: process.hrtime(),
    dnsLookupAt: undefined,
    tcpConnectionAt: undefined,
    tlsHandshakeAt: undefined,
    firstByteAt: undefined,
    endAt: undefined
  }
  const req = http.request({ ... }, (res) => {
    res.once('readable', () => {
      timings.firstByteAt = process.hrtime()
    })
    res.on('data', (chunk) => { responseBody += chunk })
    res.on('end', () => {
      timings.endAt = process.hrtime()
    })
  })
  req.on('socket', (socket) => {
    socket.on('lookup', () => {
      timings.dnsLookupAt = process.hrtime()
    })
    socket.on('connect', () => {
      timings.tcpConnectionAt = process.hrtime()
    })
    socket.on('secureConnect', () => {
      timings.tlsHandshakeAt = process.hrtime()
    })
  })

DNS Lookup only happens with domain names:

// There is no DNS lookup with IP address
const dnsLookup = dnsLookupAt !== undefined ? 
  getDuration(startAt, dnsLookupAt) : undefined

TCP Connection happens immediately after the host is resolved:

const tcpConnection = getDuration((dnsLookupAt || startAt), tcpConnectionAt)

TLS handshake (SSL) happens only with https protocol:

// There is no TLS handshake without https    
const tlsHandshake = tlsHandshakeAt !== undefined ?
      getDuration(tcpConnectionAt, tlsHandshakeAt) : undefined

We wait for server to start sending First Byte:

const firstByte = getDuration((tlsHandshakeAt || tcpConnectionAt), firstByteAt)

Content Transfer starts with the first byte:

const contentTransfer = getDuration(firstByteAt, endAt)

Total Duration is calculated from start and end dates:

const total = getDuration(startAt, endAt)

Too see the whole example together check out our https://github.com/RisingStack/example-http-timings repository.

Tools to measure timings

Now that we know how to measure HTTP timings with Node, let’s talk about existing tools that you can use to understand your HTTP requests.

request module

The popular request module has a built-in method to measure HTTP timings. You can enable it with the time property.

const request = require('request')
request({
  uri: 'https://risingstack.com',
  method: 'GET',
  time: true
}, (err, resp) => {
  console.log(err || resp.timings)
})

Distributed tracing

It’s possible to collect HTTP timings with distributed tracing tools and visualize them on a timeline. This way, you can have a full picture of what’s happening in the background and how much is the real cost of building distributed systems.

RisingStack’s opentracing-auto library has a built-in flag to collect all the HTTP timings with OpenTracing.

HTTP Request timing with opentracing-auto in Jaeger.

Summary

Measuring HTTP Timings with Node.js can help to discover performance bottlenecks. The Node ecosystem provides great tools to extract these metrics from your application.

Originally published at blog.risingstack.com on September 13, 2017.