Schedule Kumo Native App & Batch Predictions | Structured Data and Graph Models

0. Grant yourself read access to Kumo tables

Because Kumo may create new tables or update existing ones during prediction runs, grant yourself (or a reader role) access to existing and future tables in the destination schema.

Replace placeholders:

TARGET_DB / TARGET_SCHEMA → where your BP writes prediction/output tables

MY_ROLE → the role you use to read those outputs (can be ACCOUNTADMIN, or any role)

1 -- you must have these to "see" the objects
2 GRANT USAGE ON DATABASE TARGET_DB TO ROLE MY_ROLE;
3 GRANT USAGE ON SCHEMA  TARGET_DB.TARGET_SCHEMA TO ROLE MY_ROLE;
4 
5 -- read existing tables (one-time for what already exists)
6 GRANT SELECT ON ALL TABLES IN SCHEMA TARGET_DB.TARGET_SCHEMA TO ROLE MY_ROLE;
7 
8 -- read all future tables (so you don't have to grant every run)
9 GRANT SELECT ON FUTURE TABLES IN SCHEMA TARGET_DB.TARGET_SCHEMA TO ROLE MY_ROLE;

1. Define Procedures

Create two small procedures you can call from the notebook to safely start and stop the Kumo app.

Replace:

MY_DB.MY_SCHEMA → where to store these procedures (e.g., KUMO.AUTOMATION)

MY_APP → your installed app name (e.g., KUMO_SPCS_RELEASE_INTERNAL)

MY_USER_SCHEMA → schema used when installing the app

1 -- Idempotent START_APP
2 CREATE OR REPLACE PROCEDURE MY_DB.MY_SCHEMA.START_APP()
3 RETURNS STRING
4 LANGUAGE JAVASCRIPT
5 EXECUTE AS CALLER
6 AS
7 $$
8 function run(sql){ snowflake.execute({sqlText: sql}); }
9 try {
10   // If this resolves, the app is already running
11   run("CALL IDENTIFIER('MY_APP.PROCEDURES.GET_END_POINT')('MY_USER_SCHEMA')");
12   return 'App already running; no action taken.';
13 } catch (e) {
14   // Not running -> start it
15   run("CALL IDENTIFIER('MY_APP.PROCEDURES.START_APP')('MY_USER_SCHEMA')");
16   return 'App start initiated.';
17 }
18 $$;
19 
20 -- Idempotent STOP_APP
21 CREATE OR REPLACE PROCEDURE MY_DB.MY_SCHEMA.STOP_APP()
22 RETURNS STRING
23 LANGUAGE JAVASCRIPT
24 EXECUTE AS CALLER
25 AS
26 $$
27 function run(sql){ snowflake.execute({sqlText: sql}); }
28 try {
29   // Try to stop; if already stopped, the call may error — we swallow it
30   run("CALL IDENTIFIER('MY_APP.PROCEDURES.SHUTDOWN_APP')('MY_USER_SCHEMA')");
31   return 'App stop initiated.';
32 } catch (e) {
33   return 'App already stopped or stop not needed; no action taken.';
34 }
35 $$;

2. Use these in your Kumo BP SDK snippet

If you haven’t set up the SDK yet, follow the setup guide here

Rules:

Call START_APP() first (very first command).

Run your BP pipeline with non_blocking = False so the notebook waits for completion.

Call STOP_APP() last (very last command), so the app shuts down after the job.

Example notebook flow

SQL cell — start app (first line):

1 CALL MY_DB.MY_SCHEMA.START_APP();

Python cell — Kumo BP SDK (make sure non_blocking=False):

1 import kumoai
2 
3 # Gather artifacts
4 pquery = kumoai.PredictiveQuery.load_from_training_job("trainingjob-...")
5 graph = pquery.graph
6 
7 # Create a prediction table from the predictive query:
8 prediction_table = pquery.generate_prediction_table()
9 
10 # Generate predictions:
11 trainer = kumoai.Trainer.load("trainingjob-...")
12 prediction_job = trainer.predict(
13     graph,
14     prediction_table,
15     non_blocking = False  # <<< Make sure this is False so the notebook waits
16 )

SQL cell — stop app (last line):

1 CALL MY_DB.MY_SCHEMA.STOP_APP();

3) Schedule the Containerized Notebook

Open your containerized notebook in Snowflake.
Click Schedule.
Set your CRON/time zone and save

CRON examples

Every 3 hours (top of hour):
USING CRON 0 */3 * * * America/Chicago
Daily at 7:00 AM:
USING CRON 0 7 * * * America/Chicago
Daily at 8:00 PM:
USING CRON 0 20 * * * America/Chicago