Introduction
Measures time-to-first-byte for single or multiple URLs. Can show you quickest, slowest & median TTFB values plus optionally log all response headers.
Usage: ttfb [options] url [url...]
-d debug
-l <log file> (infers -d) log response headers. Defaults to ./curl.log
-n <number> of times to test time to first byte
-o <option> pass options to curl (e.g. -o "-k" will make curl ignore invalid certificates)
-v verbose output. Show response breakdown (DNS lookup, TLS handshake etc)
Implicitly follows a redirection chain using curl’s -L option.
Can log all response headers (the default log file is ./curl.log) by calling with -d.
Override the default log file by specifying -l /some/file.
Get quickest, slowest and median TTFB values by specifying the number of times to call a URL; use -n2 for 2 tests, -n5 for 5 and so on.
Uses the calculation %{time_starttransfer¹} - %{time_appconnect²} which doesn’t include any connection overhead, to better approximate devtool’s TTFB figure.
¹ time_starttransfer
The time, in seconds, it took from the start until the first byte was just about to be transferred. This includes time_pretransfer and also the time the server needed to calculate the result.
² time_appconnect
The time, in seconds, it took from the start until the SSL/SSH/etc
connect/handshake to the remote host was completed.
Genesis
Based on a gist
by https://github.com/sandeepraju
Modified by [email protected], @jaygooby
Installation
Download the script from the master branch and make it executable:
curl -LJO https://raw.githubusercontent.com/jaygooby/ttfb.sh/master/ttfb
chmod +x ./ttfb
Usage
Usage: ttfb [options] url [url...]
-d debug
-l <log file> (infers -d) log response headers. Defaults to ./curl.log
-n <number> of times to test time to first byte
-o <option> pass options to curl (e.g. -o "-k" will make curl ignore invalid certificates)
-v verbose output. Show response breakdown (DNS lookup, TLS handshake etc)
Examples
Basic usage:
$ ttfb example.com
.227436
Basic usage with verbose response breakdown:
$ ttfb -v https://example.com
DNS lookup: 0.005152 TLS handshake: 0.000000 TTFB including connection: 0.200831 TTFB: .200831 Total time: 0.201132
Test multiple times:
$ ttfb -n 5 example.com/example/url
.....
fastest .177263 slowest .214302 median .179957
Test multiple URLs:
$ ttfb bbc.co.uk news.bbc.co.uk
bbc.co.uk .049985
news.bbc.co.uk .054122
Test multiple URLs, multiple times:
$ ttfb -n 5 bbc.co.uk news.bbc.co.uk
.....
.....
bbc.co.uk fastest .030936 slowest .057755 median .034663
news.bbc.co.uk fastest .031413 slowest .182791 median .035001
Verbose response breakdown when multiple tests specified:
$ ttfb -v -n 5 bbc.co.uk
DNS lookup: 0.005335 TLS handshake: 0.102314 TTFB including connection: 0.148328 TTFB: .046014 Total time: 0.646115
DNS lookup: 0.005322 TLS handshake: 0.102609 TTFB including connection: 0.150693 TTFB: .048084 Total time: 0.644611
DNS lookup: 0.004277 TLS handshake: 0.102066 TTFB including connection: 0.172199 TTFB: .070133 Total time: 1.196256
DNS lookup: 0.004444 TLS handshake: 0.107375 TTFB including connection: 0.160771 TTFB: .053396 Total time: 0.637290
DNS lookup: 0.005352 TLS handshake: 0.118882 TTFB including connection: 0.168772 TTFB: .049890 Total time: 0.653761
fastest .046014 slowest .070133 median .049890
Log all the response headers for multiple tests to multiple URLs:
ttfb -d -n 2 bbc.co.uk https://www.bbc.co.uk/weather
..
..
bbc.co.uk fastest .027550 slowest .055215 median .041382
https://www.bbc.co.uk/weather fastest .101020 slowest .297923 median .199471
$ ls *.log
bbc_co_uk-curl.log https___www_bbc_co_uk_weather-curl.log
$ cat https___www_bbc_co_uk_weather-curl.log
HTTP/2 200
server: openresty
x-cache-action: MISS
vary: Accept-Encoding,X-BBC-Edge-Cache,X-BBC-Edge-Scheme,X-CDN
x-cache-age: 0
cache-control: private, stale-while-revalidate=10, max-age=0, must-revalidate
content-type: text/html;charset=utf-8
x-mrid: w1
date: Thu, 11 Apr 2019 17:08:07 GMT
x-xss-protection: 1; mode=block
x-content-type-options: nosniff
x-lb-nocache: true
x-msig: 24e37f81323984e4e45b8048f9e3c94a
x-frame-options: SAMEORIGIN
content-length: 1077454
HTTP/2 200
server: openresty
x-cache-action: MISS
vary: Accept-Encoding,X-BBC-Edge-Cache,X-BBC-Edge-Scheme,X-CDN
x-cache-age: 0
cache-control: private, stale-while-revalidate=10, max-age=0, must-revalidate
content-type: text/html;charset=utf-8
x-mrid: w1
date: Thu, 11 Apr 2019 17:08:08 GMT
x-xss-protection: 1; mode=block
x-content-type-options: nosniff
x-lb-nocache: true
x-msig: 24e37f81323984e4e45b8048f9e3c94a
x-frame-options: SAMEORIGIN
content-length: 1077454
More detail on time-to-first-byte
See https://blog.cloudflare.com/a-question-of-timing/
and https://curl.haxx.se/docs/manpage.html for an explanation
of how the curl variables relate to the various stages of
the transfer.
To get a better approximation of devtool’s TTFB, we consider
the time without the connection overhead:
%{time_starttransfer} - %{time_appconnect}
Uses a dirty eval to do the ttfb arithmetic. Depends
on bc and column commands.
Testing
If you’d like to submit a PR, please ensure all existing tests pass and if it adds new functionality, add a test in the test suite.
To run the tests, call make from the project root. You’ll need to have bats-core installed and in your $PATH.
If you only want to run a test or two, uncomment the # skip lines on the tests you don’t care about, so they read skip, and then those tests will be skipped.
See CONTRIBUTING.md for more details.
TODO
评论