3724 – Testing ways to better error handle RetryLimitHit (Metrics) (#394

)

* try to access the original exception for RetryLimitHit

The Exception class has a method .cause (since Ruby 2.1), with that
method we can access the original error.cause
https://ruby-doc.org/3.0.5/Exception.html

We want to see if this can help us gather more info on our exception,
however it could create duplicate data and just make it more noisy

* add config.sidekiq.report_after_job_retries = true

config/initializers/02_sentry.rb

@@ -10,4 +10,7 @@
 
   # Any exceptions we want to prevent sending to Sentry
   config.excluded_exceptions += ['Pender::Exception::RetryLater']
+
+  # report_after_job_retries when turned on, the SDK will only report the exception after all retries have failed.
+  config.sidekiq.report_after_job_retries = true
 end

config/initializers/03_sidekiq.rb

@@ -10,11 +10,11 @@
   Sidekiq.configure_server do |config|
     config.redis = redis_config
 
-    config.death_handlers << ->(job, ex) do
-      if ex.is_a?(Pender::Exception::RetryLater)
-        ex = Pender::Exception::RetryLimitHit.new(ex)
+    config.death_handlers << ->(job, original_exception) do
+      if original_exception.is_a?(Pender::Exception::RetryLater)
+        limit_hit_exception = Pender::Exception::RetryLimitHit.new(original_exception)
       end
-      PenderSentry.notify(ex, {job: job})
+      PenderSentry.notify(limit_hit_exception, {job: job, original_exception: original_exception.cause})
     end
   end