Try looping waiting for process to see if it's just a delay

spec/datadog/profiling/crashtracker_spec.rb

@@ -8,8 +8,13 @@
   before do
     skip_if_profiling_not_supported(self)
 
-    crash_tracker_pids = `pgrep -f libdatadog-crashtracking-receiver`
-    expect(crash_tracker_pids).to be_empty, "No crash tracker process should be running, found #{crash_tracker_pids}"
+    # No crash tracker process should still be running at the start of each testcase
+    wait_for { `pgrep -f libdatadog-crashtracking-receiver` }.to be_empty
+  end
+
+  after do
+    # No crash tracker process should still be running at the end of each testcase
+    wait_for { `pgrep -f libdatadog-crashtracking-receiver` }.to be_empty
   end
 
   let(:exporter_configuration) { [:agent, 'http://localhost:6006'] }