How do I force retrying Sidekiq jobs stuck in the retry queue?

Question

I have a lot of jobs clogging up the retry queue in Sidekiq, with a high latency. For some reason they are not worked on, and I cannot find a way to directly schedule the retry of these jobs from the retry queue itself.

How would I go about doing that, essentially force-retrying all of them?

Answer 1

There seems to be no way to do this directly. Retried jobs whose retries are exhausted should eventually go to the DeadSet. First collect the retried jobs and their arguments:

require 'sidekiq/api'
jobs = Sidekiq::Queue.new("retries").map { |j| [j.klass, j.args] }
# => e.g.:
#  ["MyWorker", [125396, "some_arg"]],
#  ["MyWorker", [125394, "some_arg"]],
#  ["MyWorker", [125395, "some_arg"]],
#  ["MyWorker", [98761, "another_arg"]],
#  ["MyWorker", [98760, "another_arg"]],
#  ["MyWorker", [41624, "another_arg"]],
#  ["MyWorker", [41623, "another_arg"]],
#  ["MyWorker", [32164, "another_arg"]],
#  ["MyWorker", [32167, "some_arg"]],
#  ["MyWorker", [32165, "another_arg"]],
#  ["MyWorker", [31870, "some_arg"]],
#  ["MyWorker", [31871, "some_arg"]]]

Now, schedule an immediate retry for each job:

jobs.each do |job|
  klass = job[0].constantize
  args = job[1]
  klass.perform_async(*args)
end

Finally, we clear the retry queue:

Sidekiq::Queue.new("retries").clear