Attach comments to your ActiveRecord queries. By default, it adds the application, controller, and action names as a comment at the end of each query.
This helps when searching log files for queries, and seeing where slow queries came from.
For example, once enabled, your logs will look like:
Account Load (0.3ms) SELECT `accounts`.* FROM `accounts` WHERE `accounts`.`queenbee_id` = 1234567890 LIMIT 1 /*application:BCX,controller:project_imports,action:show*/
You can also use these query comments along with a tool like pt-query-digest to automate identification of controllers and actions that are hotspots for slow queries.
This gem was created at 37signals. You can read more about how we use it on our blog.
This has been tested and used in production with the mysql2 and pg gems, and is tested on Rails 5.2 through 6.1, and Ruby 2.6 through 3.0. It is also tested for sqlite3. As of Rails 7, Marginalia is a part of Rails itself and does not need to be separately included.
Rails version support will follow supported versions in the Ruby on Rails maintenance policy and Ruby support will follow maintained versions in the Ruby maintenance policy.
Patches are welcome for other database adapters.
# Gemfile gem 'marginalia'
Optionally, you can set the application name shown in the log like so in an initializer (e.g.
Marginalia.application_name = "BCX"
The name will default to your Rails application name.
You can also configure the components of the comment that will be appended,
Marginalia::Comment.components. By default, this is set to:
Marginalia::Comment.components = [:application, :controller, :action]
Which results in a comment of
You can re-order or remove these components. You can also add additional
comment components of your desire by defining new module methods for
Marginalia::Comment which return a string. For example:
module Marginalia module Comment def self.mycommentcomponent "TEST" end end end Marginalia::Comment.components = [:application, :mycommentcomponent]
Which will result in a comment like
The calling controller is available to these methods via
Marginalia ships with
:action enabled by
default. In addition, implementation is provided for:
:line(for file and line number calling query). :line supports a configuration by setting a regexp in
Marginalia::Comment.lines_to_ignoreto exclude parts of the stacktrace from inclusion in the line comment.
:controller_with_namespaceto include the full classname (including namespace) of the controller.
:jobto include the classname of the ActiveJob being performed.
:pidto include current process id.
:db_hostto include the configured database hostname.
:socketto include the configured database socket.
:databaseto include the configured database name.
Pull requests for other included comment components are welcome.
By default marginalia appends the comments at the end of the query. Certain databases, such as MySQL will truncate the query text. This is the case for slow query logs and the results of querying some InnoDB internal tables where the length of the query is more than 1024 bytes.
In order to not lose the marginalia comments from your logs, you can prepend the comments using this option:
Marginalia::Comment.prepend_comment = true
Inline query annotations
In addition to the request or job-level component-based annotations, Marginalia may be used to add inline annotations to specific queries using a block-based API.
For example, the following code:
Marginalia.with_annotation("foo") do Account.where(queenbee_id: 1234567890).first end
will issue this query:
Account Load (0.3ms) SELECT `accounts`.* FROM `accounts` WHERE `accounts`.`queenbee_id` = 1234567890 LIMIT 1 /*application:BCX,controller:project_imports,action:show*/ /*foo*/
with_annotation blocks will concatenate the comment strings.
Be careful when using Marginalia with prepared statements. If you use a component
request_id then every query will be unique and so ActiveRecord will create
a new prepared statement for each potentially exhausting system resources.
Disable prepared statements
if you wish to use components with high cardinality values.
Start by bundling and creating the test database:
bundle rake db:mysql:create rake db:postgresql:create
rake will run the tests on all the database adapters (