Treat sidekiq nil retry as true

Author: sl0thentr0py

CHANGELOG.md

@@ -22,6 +22,8 @@
 
   If you were using the `SpanProcessor` before, we recommend migrating over to `config.otlp` since it's a much simpler setup.
 
+- Treat Sidekiq nil retry as true ([#2864](https://github.com/getsentry/sentry-ruby/pull/2864))
+
 ### Bug Fixes
 
 - Fix `MetricEvent` timestamp serialization to float ([#2862](https://github.com/getsentry/sentry-ruby/pull/2862))

sentry-sidekiq/lib/sentry/sidekiq/error_handler.rb

@@ -61,7 +61,7 @@ def call(ex, context, sidekiq_config = nil)
       def retryable?(context)
         retry_option = context.dig(:job, "retry")
         # when `retry` is not specified, it's default is `true` and it means 25 retries.
-        retry_option == true || (retry_option.is_a?(Integer) && retry_option.positive?)
+        retry_option.nil? || retry_option == true || (retry_option.is_a?(Integer) && retry_option.positive?)
       end
 
       # @return [Integer] the number of retries allowed for the job

sentry-sidekiq/spec/sentry/sidekiq_spec.rb

@@ -210,6 +210,24 @@ def retry_last_failed_job
       end
     end
 
+    context "when retry is nil in the job payload (e.g. raw payload pushed by AWS Lambda)" do
+      it "treats the job as retryable by default and does not report on first failure" do
+        context = {
+          job: {
+            "class" => "SadWorker",
+            "jid" => "abc123",
+            "queue" => "default",
+            "retry_count" => nil
+          }
+        }
+
+        handler = Sentry::Sidekiq::ErrorHandler.new
+        handler.call(RuntimeError.new("I'm sad!"), context)
+
+        expect(transport.events.count).to eq(0)
+      end
+    end
+
     context "when retry is not specified on the worker" do
       before do
         # this is required for Sidekiq to assign default options to the worker