1 Trail Demo Pipeline Setup
data-sciencenotebookData Pipeline Observabilitymachine-learningsnowflake-demo-notebooksdata-engineeringPythonsqlSnowflake Trail
Export
🍿 Snowflake Trail - Step 1: Demo Setup
Creating a dummy data ingestion and transformation project with some intentional errors and anomalies that we can monitor and alert on.
- Warehouse
- Internal Stage
- Pipe
- Dynamic Tables
- Stream and Task graph
- nested SQL & Python Procedures & Functions
Setup Details:
- 1st Task creates a new csv-file in the internal Stage
- 2nd Task creates a broken csv-file into the internal stage
- 3rd Task manually refreshes the Pipe to load the 2 new files from the source Stage into a target table
- Pipe copy of the broken csv-file will fail -> logging file ingestion error to the event table
- the Stream on the target table will catch the new data from the successful file ingestion
- Stream triggers the root Task to run the graph
- child Tasks call procedures and functions, which may fail at random -> logging Procedure and Task errors
- parallel child Task will change the name of the source table of a Dynamic Table and then manually refresh the Dynamic Table -> logging Dynamic Table refresh error
-> manually run the Trigger Task via the UI or SQL to create new logs
[ ]
[ ]
[ ]
[ ]
[ ]
[ ]
[ ]
[ ]
[ ]
[ ]
[ ]
[ ]
[ ]
[ ]
[ ]
[ ]
[ ]
[ ]
[ ]
[ ]
[ ]
[ ]
[ ]
[ ]
[ ]
[ ]
[ ]
[ ]
[ ]
[ ]
[ ]
Now you can intentionally trigger a range of event logs from
Task run, Pipe copy, Dynamic Table refresh, Snowpark Procedure and Function calls
to test or demo the logging to the event table.
[ ]
continue with the
Snowflake Trail - Step 2: Observability Setup Notebook
to add a comprehensive observability-layer to this pipeline.