datus-ai EMR Serverless
datus emr-serverless operates EMR Serverless applications and Spark job runs
through boto3. Global usage:
datus emr-serverless [--profile <env>] <group> <subcommand> [args...]
Add -o json for full output. An application must be STARTED before it can
run jobs. jobs run submits a Spark job (writes data, billed).
Applications
datus emr-serverless applications list [--limit N]
datus emr-serverless applications get <app-id>
datus emr-serverless applications start <app-id> # warm capacity so it can run jobs
datus emr-serverless applications stop <app-id>
Jobs
datus emr-serverless jobs run [<app-id>] --entry-point s3://bkt/job.py \
[--entry-point-args '["--date","2026-07-01"]'] [--spark-submit-params "--conf spark.executor.memory=4g"] \
[--execution-role arn:...] [--name my-run] [--wait]
datus emr-serverless jobs list <app-id> [--limit N]
datus emr-serverless jobs run-status <app-id> <run-id>
datus emr-serverless jobs cancel <app-id> <run-id>
datus emr-serverless jobs dashboard <app-id> <run-id> # live Spark UI URL
<app-id> and --execution-role default to the profile's application_id /
execution_role_arn. run --wait polls until the run is terminal: exit 0 on
SUCCESS, 1 otherwise. dashboard returns the live Spark UI URL (use it to
inspect logs and stages).
Exit codes
0 success · 1 runtime/API error (also: failed run under --wait) · 2
usage (also: missing application id / execution role) · 3 config error.