Middleware that displays speed badge for every html page. Designed to work both in production and in development.
- Database profiling - Currently supports Mysql2, Postgres, Oracle (oracle_enhanced ~> 1.5.0) and Mongoid3 (with fallback support to ActiveRecord)
- Call-stack profiling - Flame graphs showing time spent by gem
- Memory profiling - Per-request memory usage, GC stats, and global allocation metrics
- Visit our community
- Watch the RailsCast
- Read about Flame graphs in rack-mini-profiler
- Read the announcement posts from 2012
We have decided to restructure our repository so there is a central UI repo and the various language implementation have their own.
WE NEED HELP.
- Setting up a build that reuses https://github.com/MiniProfiler/ui
- Migrating the internal data structures per the spec
If you feel like taking on any of this start an issue and update us on your progress.
Install/add to Gemfile
gem 'rack-mini-profiler'
NOTE: Be sure to require rack_mini_profiler below the pg
and mysql
gems in your Gemfile. rack_mini_profiler will identify these gems if they are loaded to insert instrumentation. If included too early no SQL will show up.
You can also include optional libraries to enable additional features.
# For memory profiling (requires Ruby MRI 2.1+)
gem 'memory_profiler'
# For call-stack profiling flamegraphs (requires Ruby MRI 2.0.0+)
gem 'flamegraph'
gem 'stackprof' # For Ruby MRI 2.1+
gem 'fast_stack' # For Ruby MRI 2.0
All you have to do is to include the Gem and you're good to go in development. See notes below for use in production.
In case you need to make sure rack_mini_profiler initialized is after all other gems, or you want to execute some code before rack_mini_profiler required:
gem 'rack-mini-profiler', require: false
Note the require: false
part - if omitted, it will cause the Railtie for the mini-profiler to
be loaded outright, and an attempt to re-initialize it manually will raise an exception.
Then put initialize code in file like config/initializers/rack_profiler.rb
if Rails.env == 'development'
require 'rack-mini-profiler'
# initialization is skipped so trigger it
Rack::MiniProfilerRails.initialize!(Rails.application)
end
require 'rack-mini-profiler'
home = lambda { |env|
[200, {'Content-Type' => 'text/html'}, ["<html><body>hello!</body></html>"]]
}
builder = Rack::Builder.new do
use Rack::MiniProfiler
map('/') { run home }
end
run builder
require 'rack-mini-profiler'
class MyApp < Sinatra::Base
use Rack::MiniProfiler
end
For working with hanami, you need to use rack integration. Also, you need to add Hanami::View::Rendering::Partial#render
method for profile:
# config.ru
require 'rack-mini-profiler'
Rack::MiniProfiler.profile_method(Hanami::View::Rendering::Partial, :render) { "Render partial #{@options[:partial]}" }
use Rack::MiniProfiler
A typical web application spends a lot of time querying the database. rack_mini_profiler will detect the ORM that is available and apply patches to properly collect query statistics.
To make this work, declare the orm's gem before declaring rack-mini-profiler
in the Gemfile
:
gem 'pg'
gem 'mongoid'
gem 'rack-mini-profiler'
If you wish to override this behavior, the environment variable RACK_MINI_PROFILER_PATCH
is available.
export RACK_MINI_PROFILER_PATCH="pg,mongoid"
# or
export RACK_MINI_PROFILER_PATCH="false"
# initializers/rack_profiler.rb: SqlPatches.patch %w(mongo)
To generate flamegraphs:
- add the flamegraph gem to your Gemfile
- visit a page in your app with
?pp=flamegraph
Flamegraph generation is supported in Ruby MRI 2.0+
Memory allocations can be measured (using the memory_profiler gem)
which will show allocations broken down by gem, file location, and class and will also highlight String
allocations.
(Requires Ruby MRI 2.1.0+)
Add ?pp=profile-memory
to the URL of any request while Rack::MiniProfiler is enabled to generate the report.
Additional query parameters can be used to filter the results.
memory_profiler_allow_files
- filename pattern to include (default is all files)memory_profiler_ignore_files
- filename pattern to exclude (default is no exclusions)memory_profiler_top
- number of results per section (defaults to 50)
The allow/ignore patterns will be treated as regular expressions.
Example: ?pp=profile-memory&memory_profiler_allow_files=active_record|app
There are two additional pp
options that can be used to analyze memory which do not require the memory_profiler
gem
- Use
?pp=profile-gc
to report on Garbage Collection statistics (requires Ruby MRI 1.9.3+) - Use
?pp=analyze-memory
to report on ObjectSpace statistics (requires Ruby 2.0.0+)
rack-mini-profiler is designed with production profiling in mind. To enable that just run Rack::MiniProfiler.authorize_request
once you know a request is allowed to profile.
# inside your ApplicationController
before_action do
if current_user && current_user.is_admin?
Rack::MiniProfiler.authorize_request
end
end
Various aspects of rack-mini-profiler's behavior can be configured when your app boots. For example in a Rails app, this should be done in an initializer: config/initializers/mini_profiler.rb
To fix some nasty bugs with rack-mini-profiler showing the wrong data, the middleware
will remove headers relating to caching (Date & Etag on responses, If-Modified-Since & If-None-Match on requests).
This probably won't ever break your application, but it can cause some unexpected behavior. For
example, in a Rails app, calls to stale?
will always return true.
To disable this behavior, use the following config setting:
# Do not let rack-mini-profiler disable caching
Rack::MiniProfiler.config.disable_caching = false # defaults to true
rack-mini-profiler stores its results so they can be shared later and aren't lost at the end of the request.
There are 4 storage options: MemoryStore
, RedisStore
, MemcacheStore
, and FileStore
.
FileStore
is the default in Rails environments and will write files to tmp/miniprofiler/*
. MemoryStore
is the default otherwise.
# set MemoryStore
Rack::MiniProfiler.config.storage = Rack::MiniProfiler::MemoryStore
# set RedisStore
if Rails.env.production?
uri = URI.parse(ENV["REDIS_SERVER_URL"])
Rack::MiniProfiler.config.storage_options = { :host => uri.host, :port => uri.port, :password => uri.password }
Rack::MiniProfiler.config.storage = Rack::MiniProfiler::RedisStore
end
MemoryStore
stores results in a processes heap - something that does not work well in a multi process environment.
FileStore
stores results in the file system - something that may not work well in a multi machine environment.
RedisStore
/MemcacheStore
work in multi process and multi machine environments (RedisStore
only saves results for up to 24 hours so it won't continue to fill up Redis). You will need to add gem redis
/gem dalli
respectively to your Gemfile
to use these stores.
Additionally you may implement an AbstractStore
for your own provider.
MiniProfiler will attempt to keep all user results isolated, out-of-the-box the user provider uses the ip address:
Rack::MiniProfiler.config.user_provider = Proc.new{|env| Rack::Request.new(env).ip}
You can override (something that is very important in a multi-machine production setup):
Rack::MiniProfiler.config.user_provider = Proc.new{ |env| CurrentUser.get(env) }
The string this function returns should be unique for each user on the system (for anonymous you may need to fall back to ip address)
You can increase the granularity of profiling by measuring the performance of specific methods. Add methods of interest to an initializer.
Rails.application.config.to_prepare do
::Rack::MiniProfiler.profile_singleton_method(User, :non_admins) { |a| "executing all_non_admins" }
::Rack::MiniProfiler.profile_method(User, :favorite_post) { |a| "executing favorite_post" }
end
It is also possible to profile any arbitrary block of code by passing a block to Rack::MiniProfiler.step(name, opts=nil)
.
Rack::MiniProfiler.step('Adding two elements') do
result = 1 + 2
end
Single page applications built using Ember, Angular or other frameworks need some special care, as routes often change without a full page load.
On route transition always call:
window.MiniProfiler.pageTransition();
This method will remove profiling information that was related to previous page and clear aggregate statistics.
You need to inject the following in your SPA to load MiniProfiler's speed badge (extra details surrounding this script):
<script async type="text/javascript" id="mini-profiler" src="/mini-profiler-resources/includes.js?v=12b4b45a3c42e6e15503d7a03810ff33" data-version="12b4b45a3c42e6e15503d7a03810ff33" data-path="/mini-profiler-resources/" data-current-id="redo66j4g1077kto8uh3" data-ids="redo66j4g1077kto8uh3" data-position="left" data-trivial="false" data-children="false" data-max-traces="10" data-controls="false" data-authorized="true" data-toggle-shortcut="Alt+P" data-start-hidden="false" data-collapse-results="true"></script>
Note: The GUID (data-version
and the ?v=
parameter on the src
) will change with each release of rack_mini_profiler
. The MiniProfiler's speed badge will continue to work, although you will have to change the GUID to expire the script to fetch the most recent version.
You can set configuration options using the configuration accessor on Rack::MiniProfiler
.
For example:
Rack::MiniProfiler.config.position = 'right'
Rack::MiniProfiler.config.start_hidden = true
The available configuration options are:
Option | Default | Description |
---|---|---|
pre_authorize_cb | Rails: dev only Rack: always on |
A lambda callback that returns true to make mini_profiler visible on a given request. |
position | 'left' |
Display mini_profiler on 'right' or 'left' . |
skip_paths | [] |
Paths that skip profiling. |
skip_schema_queries | Rails dev: 'true' Othwerwise: 'false' |
'true' to log schema queries. |
auto_inject | true |
true to inject the miniprofiler script in the page. |
backtrace_ignores | [] |
Regexes of lines to be removed from backtraces. |
backtrace_includes | Rails: `[/^/?(app | config |
backtrace_remove | rails: Rails.root Rack: nil |
A string or regex to remove part of each line in the backtrace. |
toggle_shortcut | Alt+P | Keyboard shortcut to toggle the mini_profiler's visibility. See jquery.hotkeys. |
start_hidden | false |
false to make mini_profiler visible on page load. |
backtrace_threshold_ms | 0 |
Minimum SQL query elapsed time before a backtrace is recorded. Backtrace recording can take a couple of milliseconds on rubies earlier than 2.0, impacting performance for very small queries. |
flamegraph_sample_rate | 0.5 |
How often to capture stack traces for flamegraphs in milliseconds. |
disable_env_dump | false |
true disables ?pp=env , which prevents sending ENV vars over HTTP. |
base_url_path | '/mini-profiler-resources/' |
Path for assets; added as a prefix when naming assets and sought when responding to requests. |
collapse_results | true |
If multiple timing results exist in a single page, collapse them till clicked. |
max_traces_to_show | 20 | Maximum number of mini profiler timing blocks to show on one page |
html_container | body |
The HTML container (as a jQuery selector) to inject the mini_profiler UI into |
If you are using Rack::Deflate
with rails and rack-mini-profiler in its default configuration,
Rack::MiniProfiler
will be injected (as always) at position 0 in the middleware stack. This
will result in it attempting to inject html into the already-compressed response body. To fix this,
the middleware ordering must be overriden.
To do this, first add , require: false
to the gemfile entry for rack-mini-profiler.
This will prevent the railtie from running. Then, customize the initialization
in the initializer like so:
require 'rack-mini-profiler'
Rack::MiniProfilerRails.initialize!(Rails.application)
Rails.application.middleware.delete(Rack::MiniProfiler)
Rails.application.middleware.insert_after(Rack::Deflater, Rack::MiniProfiler)
Deleting the middleware and then reinserting it is a bit inelegant, but a sufficient and costless solution. It is possible that rack-mini-profiler might support this scenario more directly if it is found that there is significant need for this confriguration or that the above recipe causes problems.
If you include the query string pp=help
at the end of your request you will see the various options available. You can use these options to extend or contract the amount of diagnostics rack-mini-profiler gathers.
To get MiniProfiler working with Rails 2.3.X you need to do the initialization manually as well as monkey patch away an incompatibility between activesupport and json_pure.
Add the following code to your environment.rb (or just in a specific environment such as development.rb) for initialization and configuration of MiniProfiler.
# configure and initialize MiniProfiler
require 'rack-mini-profiler'
c = ::Rack::MiniProfiler.config
c.pre_authorize_cb = lambda { |env|
Rails.env.development? || Rails.env.production?
}
tmp = Rails.root.to_s + "/tmp/miniprofiler"
FileUtils.mkdir_p(tmp) unless File.exist?(tmp)
c.storage_options = {:path => tmp}
c.storage = ::Rack::MiniProfiler::FileStore
config.middleware.use(::Rack::MiniProfiler)
::Rack::MiniProfiler.profile_method(ActionController::Base, :process) {|action| "Executing action: #{action}"}
::Rack::MiniProfiler.profile_method(ActionView::Template, :render) {|x,y| "Rendering: #{path_without_format_and_extension}"}
# monkey patch away an activesupport and json_pure incompatability
# http://pivotallabs.com/users/alex/blog/articles/1332-monkey-patch-of-the-day-activesupport-vs-json-pure-vs-ruby-1-8
if JSON.const_defined?(:Pure)
class JSON::Pure::Generator::State
include ActiveSupport::CoreExtensions::Hash::Except
end
end
$ rake build
$ rake spec
Additionally you can also run autotest
if you like.
The MIT License (MIT)
Copyright (c) 2013 Sam Saffron
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.