[EBPF] kmt: avoid cleanup job failures on canceled pipelines (#33525)

Author: gjulianm

.gitlab/kernel_matrix_testing/common.yml

@@ -218,7 +218,7 @@
     # setup_env job hasn't finished. This causes instances to be leftover for more time than necessary.
     - inv kmt.wait-for-setup-job --pipeline-id $CI_PIPELINE_ID --arch $ARCH --component $TEST_COMPONENT
     - aws ec2 describe-instances --filters $FILTER_TEAM $FILTER_MANAGED $FILTER_PIPELINE $FILTER_ARCH $FILTER_INSTANCE_TYPE $FILTER_TEST_COMPONENT --output json --query $QUERY_INSTANCE_IDS | tee -a instance.json
-    - cat instance.json | jq -r 'map(.[]) | .[]' | grep -v "null" | xargs -n 1 -t aws ec2 terminate-instances --instance-ids
+    - cat instance.json | jq -r 'map(.[]) | .[]' | grep -v "null" | xargs --no-run-if-empty -n 1 -t aws ec2 terminate-instances --instance-ids
   after_script:
     - DD_API_KEY=$($CI_PROJECT_DIR/tools/ci/fetch_secret.sh $AGENT_API_KEY_ORG2 token) || exit $?; export DD_API_KEY
     - !reference [.tag_kmt_ci_job]