Things to Remember When Working with Sidekiq

Photo of Wojtek Zagórski

Wojtek Zagórski

Updated Feb 27, 2023 • 3 min read

Even though the free version of Sidekiq lacks a few useful features, it’s still the most popular background processing gem for Ruby on Rails apps thanks to its performance and reliability.

In Netguru, it’s mostly used for time-consuming tasks such as sending emails to customers or long database queries. There are few things you must remember about when working with Sidekiq.

TIP: Keep Jobs Parameters Simple

Sidekiq jobs can receive parameters of any format, but the more complex they are, the more likely they are to become corrupted. The parameters are serialised and stored in Redis until the queued job is run. It’s a good practice to pass only the most basic data types: integers (database IDs) or strings.

TIP: Don’t Expect Jobs to Be Completed in Fixed Time

It’s essential to design the application independent of the jobs' completion time. Even when jobs are mostly run in two seconds, they may always get enqueued for a long time, especially if the application scales up.


TIP: Binding.pry Works with Sidekiq in Development Environment

Even though the jobs are run in different threads, the binding.pry works fine. Developers tend to think that it doesn’t because of lack of code autoreload. You need to restart Sidekiq after every change in the code.


TIP: Remember about Redis Namespaces

When Redis is shared with many containers or apps, Sidekiq needs to be configured to use namespaces to avoid running jobs from other applications. You can use a gem called "redis-namespace" and setup a proper namespace in config/initializers/sidekiq.rb:

Sidekiq.configure_server do |config|
config.redis = { url: AppConfig.redis.url, namespace: AppConfig.redis_namespace }
end

Sidekiq.configure_client do |client|
client.redis = { url: AppConfig.redis.url, namespace: AppConfig.redis_namespace }
end

TIP: Remember to Clear the Retries Queue

Don’t forget about jobs that fail during the development. They are put in the retires queue and rerun every few minutes, even after you fix your code. By spamming the Sidekiq output with continuous failures, they may make you think that the job doesn’t work properly. Remember to clear them in /sidekiq/retries.


I hope these few tips make your dev’s life easier. Take care!

Photo of Wojtek Zagórski

More posts by this author

Wojtek Zagórski

Wojtek graduated from Poznan University of Technology earning a Master's degree in Information...
New call-to-action

We're Netguru!

At Netguru we specialize in designing, building, shipping and scaling beautiful, usable products with blazing-fast efficiency
Let's talk business!

Trusted by: