GCP Integration Tests Catch Failures (#580)

## Summary
Right now, the gcp integration tests do not recognize failures from
dataproc and continue running the commands. This has led to the
group_by_upload job hanging indefinitely. This change ensures that if
the dataproc job state is "ERROR" we fail immediately.

## Checklist
- [ ] Added Unit Tests
- [ ] Covered by existing CI
- [ ] Integration tested
- [ ] Documentation update



<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->

## Summary by CodeRabbit

- **Bug Fixes**
- Enhanced job status checking by ensuring that both empty and
explicitly error-indicating states trigger consistent failure responses,
resulting in more robust job evaluation.

<!-- end of auto-generated comment: release notes by coderabbit.ai -->

Author: chewy-zlai

scripts/distribution/run_gcp_quickstart.sh

@@ -55,10 +55,10 @@ function check_dataproc_job_state() {
         exit 1
   fi
   echo -e "${GREEN} <<<<<<<<<<<<<<<<-----------------JOB STATUS----------------->>>>>>>>>>>>>>>>>\033[0m"
-  JOB_STATE=$(gcloud dataproc jobs describe $JOB_ID --region=us-central1 --format=flattened | grep "status.state:")
+  JOB_STATE=$(gcloud dataproc jobs describe $JOB_ID --region=us-central1 --format=flattened | grep "status.state:" | awk '{print $NF}')
   echo $JOB_STATE
 #  TODO: this doesn't actually fail. need to fix.
-  if [ -z "$JOB_STATE" ]; then
+  if [ -z "$JOB_STATE" ] || [ "$JOB_STATE" == "ERROR" ]; then
         echo -e "${RED} <<<<<<<<<<<<<<<<-----------------JOB FAILED!----------------->>>>>>>>>>>>>>>>>\033[0m"
         exit 1
   fi