[ferries] Run datakit smoke at production priority

Match the canary ferry pattern so the daily smoke run schedules
ahead of researcher jobs and is not preempted, keeping the lane
representative of intended hardware availability.

Author: ravwojdyla

.github/workflows/marin-datakit-smoke.yaml

@@ -69,6 +69,7 @@ jobs:
           JOB_ID=$(.venv/bin/iris --config=${{ env.IRIS_CONFIG }} \
             job run --no-wait \
             --memory=2G --disk=4G --cpu=1 --extra=cpu \
+            --priority production \
             -e SMOKE_RUN_ID "$SMOKE_RUN_ID" \
             -e FERRY_STATUS_PATH "$FERRY_STATUS_PATH" \
             -e WANDB_ENTITY "$WANDB_ENTITY" \