Warehouse Validation

suky is zero-ingestion. Every quality metric — freshness, volume, null rate, schema compliance, SLA status — is computed directly in your Snowflake warehouse via read-only SQL, and only the computed results are persisted in suky.

How it works

1. You connect a least-privilege Snowflake role (USAGE on warehouse/database/schema, SELECT on event tables).
2. The snowflake-discover edge function inventories candidate event tables and columns.
3. The snowflake-validate and snowflake-observe functions run incremental SQL to produce metrics and violation diffs against your active contracts.
4. suky stores computed results — counts, rates, per-event health scores, violation records — plus a small sample of values used for discovery. Your raw event stream never leaves your warehouse.

What gets computed

• Whether the event exists in the event schema.
• Missing required properties (null rates).
• Type mismatches and schema violations.
• Volume trends, freshness, and SLA compliance.
• Anomalies: flatlines, volume spikes/drops vs baseline, rising violation rates.

Privacy invariant

suky has no endpoints that accept raw event payloads. There are no webhook URLs for CDPs to ship events to. Any previous ingest endpoints (ingest-segment, ingest-webhook) have been retired and return 410 Gone.

Schema registry enforcement

To enforce contracts at ingestion time — before events reach your warehouse — push them to your CDP's schema registry (Segment Protocols, RudderStack Tracking Plans, Snowplow Iglu, mParticle Data Master). See Schema Registry Publishing.