• Stars
    star
    20
  • Rank 1,121,974 (Top 23 %)
  • Language
    Ruby
  • License
    MIT License
  • Created over 5 years ago
  • Updated 18 days ago

Reviews

There are no reviews yet. Be the first to send feedback to the community and the maintainers!

Repository Details

Automate routines before/after deployment

License: MIT Gem Version Verification Publish Gem

DeployPin

DeployPin

Deploying applications often involves the need to execute a series of specific commands or tasks either before or after the deployment process. While you might typically turn to migrations for such operations, these tasks aren't always migration-related. Additionally, there are situations where you need to execute code before or after a migration without causing the main thread to block.

Introducing deploy_pins โ€“ your go-to solution for streamlined task management during the deployment process. This Ruby library allows you to seamlessly orchestrate tasks before, after, or independently of migrations, offering the flexibility you need to maintain a smooth and efficient deployment workflow. With deploy_pins, you can take control of your deployment tasks and ensure that your application operates flawlessly in any environment.

Usage

DeployPin

To generate a new task template file:

rails g deploy_pin:task some_task_title
# or
rails g deploy_pin:task some_task_title --parallel

You can also specify the author:

rails g deploy_pin:task some_task_title -a author_name

To list all pending tasks:

rake deploy_pin:list

To run all pending tasks:

rake deploy_pin:run

Grouped Tasks

To define allowed groups, navigate to config/initializers/deploy_pin.rb. You can group tasks around the "allowed_group" like this:

rails g deploy_pin:task task_title -g allowed_group
# or
rails g deploy_pin:task task_title -g allowed_group --parallel

To list all pending tasks within the "allowed_group":

rake deploy_pin:list[allowed_group]

To run all pending tasks within the "allowed_group":

rake deploy_pin:run[allowed_group]

Run by Identifier

To execute a specific task using its identifier:

rake deploy_pin:run['identifier_1, identifier_2']

Alternatively, you can combine an identifier and a group:

rake deploy_pin:run['identifier, allowed_group']

If you wish to rerun a task, add an exclamation mark at the end of the identifier:

rake deploy_pin:run['identifier_1!, identifier_2!']

Installation

Add this line to your application's Gemfile:

gem 'deploy_pin'

Then execute:

$ bundle

You can also install it manually with:

$ gem install deploy_pin

Afterward, generate the configuration file:

rails g deploy_pin:install

Finally, run the migration:

rake db:migrate

Database Timeout

By default, deploy_pin runs all non-parallel tasks under a database statement timeout. To set a default value, you should define it in the deploy_pin initializer, for example:

# config/initializers/deploy_pin.rb
DeployPin.setup do
  statement_timeout 0.2.seconds # 200 ms
end

If you want to use a different value than the default, you need to specify it explicitly in the task, as shown below:

# Some deploy_pin task
# 20190401135040:I
# task_title: Execute some query with timeout

# === task code goes down here ===
DeployPin::Database::execute_with_timeout do
 ActiveRecord::Base.connection.execute("select * from shipments;")
end

For more information about the parameters, please refer to the documentation here.

Parallel

To run parallel tasks using a timeout, you need to use the parallel wrapper, which mimics the parallel interface but adds the timeout option. In a deploy_pin task, instead of using Parallel.each(1..2, in_processes: 2), use:

parallel_each(1..2, in_processes: 2, timeout: 0.3.seconds) do |i|
  # ActiveRecord::Base.connection_pool.with_connection is already included in the parallel wrapper.
  puts "Item: #{i}, Worker: #{Parallel.worker_number}"
  ActiveRecord::Base.connection.execute("<some db query>")
end

Check the documentation here for more details.

Formatting

run_formatter is used to format the output of a run task, and list_formatter is used to format the output of a list task. To set a default value, you should define it in the deploy_pin initializer:

# config/initializers/deploy_pin.rb
DeployPin.setup do
  run_formatter(
    lambda do |index, task_count, task, executable, start, duration = nil|
      end_of_msg = if executable
                     start ? '(Started)' : "(Done in #{duration})\n\n"
                   else
                     "(Skipped)\n\n"
                   end

      puts("[#{index + 1}/#{task_count}] Task #{task.title} #{task.identifier}##{task.group} #{end_of_msg}".blue.bold)
    end
  )
  list_formatter(
    lambda do |index, task|
      puts("======= Task ##{index} ========".blue.bold)

      # Print details
      task.details.each do |key, value|
        puts("#{key}:\t\t#{value}")
      end

      puts("\n<<<\n#{task.script.strip.green}\n>>>\n\n")
    end
  )
end

To use a different formatting value than the default, you need to specify it explicitly in the task, similar to the database timeout configuration.

Recurring Tasks

If you want to generate a recurring task, you can use the --recurring option. Make sure to set a correct --identifier, which should be a numeric value. Positive and negative numbers are possible here. The identifier affects the order of task execution, allowing you to customize the sequence as desired.

Please note that two identifiers, 0 and -10, are already reserved for deployment state tracking. Avoid using these identifiers.

rails g deploy_pin:task some_task_title --recurring --identifier 5
# or
rails g deploy_pin:task some_task_title --parallel --recurring --identifier 5

DeploymentStateTrack

In the initializer

    DeployPin.setup do
      groups %w[I II III post rollback]
      ...
      deployment_state_transition({
        ongoing: %w[I III],
        pending: "rollback", # enters to pending step before "rollback"
        ttl: 20.second, # memoize the state to avoid Redis spam
        redis_url: "redis://localhost:6379"
      })
    end

    # enabled next methods
    DeployPin.ongoing_deployment?
    DeployPin.pending_deployment?

Around the deployment

    bundle exec rake deploy_pin:run[I, II, III] - # enters to ongoing state before "I" and leaves it after "III" so all tasks in I, II, III have DeployPin.oingoing_deployment? == true
    bundle exec rake deploy_pin:run[rollback] - # enters "pending state"

Contributing

Contribution directions go here.

License

The gem is available as open source under the terms of the MIT License.