Authoring Schemas
What are Schemas
Snowplow schemas are self-describing JSONs used to add additional context to events. Think of Google Analytics custom dimensions or Adobe Traffic Variables but nicely grouped under a logical name.
The grouping is essential as it provides a layer of governance, unlike sending Adhoc JSON payloads like other vendors.
Snowplow also enforces that the payload matches the schema specifications; otherwise, the event will be considered invalid.
There are several advantages:
- You will never run out of space for context
- Schemas are versioned and can expand/contract with business needs
- Schemas are reusable
- Schema variables are easy to read and understand
- Easier data modeling
WARNING
Schemas are immutable. Once schemas are in production, they can't be edited or deleted. To alter a schema, you need to increment its version. This ensures compatibility with existing properties that might be sending events with existing schemas.
Authoring Schemas
- Download the template schema boilerplate and rename it to iglu-yourcompany-schema-registry.
- Version your Iglu schema registry with git
- Install igluctl
- Use a code editor to author schema (e.g. VSCode)
- Validate schemas with igluctl and Snowplow Micro
Best Practices
- Always validate your schemas with igluctl
- If your schema has a verb name you're doing it wrong (e.g. feature_click)