Resque showing job as processed despite raised exception in Exq

[eidge]

Hey there guys,

I'm using Exq to process some of our resque jobs in elixir. Everything is working and I'm able to enqueue jobs on the Ruby side and process them in the Elixir side with Exq.

The only problem is that I'm forcing my job to raise an exception in some cases and was expecting the failed job to show on Resque, but it just shows a successfully processed job instead.

My job looks like this:

defmodule SendMessage do
  @moduledoc """
  Exq worker to consume notification requests from redis.
  """

  alias CherAmi.MessageHandler
  alias CherAmiWeb.Mappers.MessageMapper

  def perform(message) do
    result = message
    |> MessageMapper.map
    |> MessageHandler.notify

    case result do
      {:ok, message} -> true
      {:error, errors} ->
        raise(ArgumentError, "Message is invalid. Message: #{inspect message}")
    end
  end
end

And Exq acknowledges the raised exception:

[error] Process #PID<0.481.0> raised an exception
** (ArgumentError) Message is invalid. Message: %{"data" => %{"key" => "created", "resource_id" => 251911, "resource_type" => "alarm", "text" => "New high warning for Simões"}, "user_ids" => [1366, 1624, 1626, 1767, 1633, 1632, 1634, 1635, 1623, 1636, 1620]}
    (cher_ami_redis) lib/send_message.ex:19: SendMessage.perform/1
    (exq) lib/exq/worker/server.ex:119: anonymous fn/3 in Exq.Worker.Server.dispatch_work/2

But resque still shows this as a success.

Am I doind something wrong or is this a bug?

[eidge]

Actually even Exq.Api.stats reports all as processed and none as failed.

[akira]

@eidge Please note that by default, failed items will be retried. You can change this in config though to 0 retries, which would then be counted as failed after one error.

See this session below:

iex(2)> 

defmodule MyTest do
   def perform do
      raise(ArgumentError, "Message is invalid. Message:")
   end
end

iex(17)> Exq.enqueue(Exq, "default", MyTest, [])
{:ok, "4d6ab221-55ca-40a8-b88b-dcf9a3942d57"}
iex(18)>
2017-03-07 20:52:59.311 [info]: Elixir.MyTest[4d6ab221-55ca-40a8-b88b-dcf9a3942d57] start

2017-03-07 20:52:59.315 [error]: Process #PID<0.467.0> raised an exception
** (ArgumentError) Message is invalid. Message:
    iex:18: MyTest.perform/0
    (exq) lib/exq/worker/server.ex:119: anonymous fn/3 in Exq.Worker.Server.dispatch_work/2

2017-03-07 20:52:59.315 [info]: Queueing job 4d6ab221-55ca-40a8-b88b-dcf9a3942d57 to retry in 36.0 seconds

2017-03-07 20:52:59.316 [info]: {%ArgumentError{message: "Message is invalid. Message: "},
 [{MyTest, :perform, 0, [file: 'iex', line: 18]},
  {Exq.Worker.Server, :"-dispatch_work/2-fun-0-", 3,
   [file: 'lib/exq/worker/server.ex', line: 119]}]}

2017-03-07 20:52:59.316 [info]: Elixir.MyTest[4d6ab221-55ca-40a8-b88b-dcf9a3942d57] fail: 5ms sec

iex(21)> Exq.Api.retries(Exq.Api)
{:ok,
 [%Exq.Support.Job{args: [], class: "MyTest", enqueued_at: 1488948779.297586,
   error_class: nil,
   error_message: "{%ArgumentError{message: \"Message is invalid. Message: \"},\n [{MyTest, :perform, 0, [file: 'iex', line: 18]},\n  {Exq.Worker.Server, :\"-dispatch_work/2-fun-0-\", 3,\n   [file: 'lib/exq/worker/server.ex', line: 119]}]}",
   failed_at: 1488948909.748436, finished_at: nil,
   jid: "4d6ab221-55ca-40a8-b88b-dcf9a3942d57", processor: nil,
   queue: "default", retry: 25, retry_count: 3}]}

It should be in the "retries" list as above. Is this happening to you also? If not, maybe can you give a script or file to replicate it?

[eidge]

Hey there @akira, I have set the retries to 0 actually.

I'll try to create a reproducible example tomorrow.