Data quality monitoring that lives inside Snowflake
Data Shepherd is a Snowflake Native App. You point it at tables you already have, set rules against them, and it records what it finds - all inside your own account, with zero data egress.
25 built-in rules, in five families.
You choose a table and column, fill in a form, and run the rule by hand or on a schedule. Every rule is read-only: none of them writes to, updates or deletes a row in the data you monitor.
Core checks
Completeness, uniqueness, value range, data freshness and referential integrity, plus table profiling and an overall quality score.
Validation checks
Allowed-value domains, fields that must be filled when a condition is met, text hiding in numeric columns, and near-duplicate detection.
Historical checks
Distribution drift, period-on-period comparison, trend, seasonal and late-data checks that compare each run against your own history.
Pattern checks
Regular-expression patterns, pattern comparison across segments, time-series pattern detection and density clustering.
Schema drift
Each connected table's columns, types and nullability are snapshotted. You are alerted when a column is added, dropped or retyped.
Suggested rules
Not sure where to start? Data Shepherd can profile a connected table and suggest starter rules for it, such as completeness, uniqueness, value range and freshness checks. You review them and apply the ones you want.
One place to set rules and see what they found.
The app opens in Snowsight, alongside the rest of your Snowflake work. There is no separate login and nothing extra to host.
Dashboard
Pass rate, active rules, open alerts and run counts over 24 hours, 7 days or 30 days, with a pass-rate score for each quality dimension your rules cover.
Rules
Create, edit, run and disable rules from a form. Set where each rule's alerts go while you are there.
Executions and alerts
Every run with its status, counts and full result. Every alert a rule raised, and who acknowledged it.
Schedules
Which rules are scheduled, when each next runs, and a pause for any schedule you want to hold.
Reports
17 built-in reports. Each downloads as CSV, JSON or PDF, and a summary can be emailed to your team daily or weekly.
Settings and setup
Connect tables, manage alert contacts, set quiet hours and choose how long run results, alerts and logs are kept.
Told when it matters, where you already work.
Alerts are delivered through Snowflake's own notification service, from your account. Data Shepherd makes no outbound network call of its own.
Email or Slack
Send alerts by email to people in your Snowflake account, or to a Slack channel. Each rule chooses its own email recipients, and whether it also posts to your Slack channel.
Acknowledge and audit
Acknowledge an alert in the app and any message not yet sent is cancelled. Who acknowledged what, and when, is kept alongside every rule change and run.
Quiet hours, fewer messages
Quiet hours hold alert delivery during a window you choose, without pausing the rules, and deliver what was held when the window ends. Alerts on the same table raised in the same five-minute window arrive as one message.
See a rule run before you connect anything.
You do not need to point Data Shepherd at your own data to see it work.
Built-in sample data
One button loads two small sample tables with deliberate quality problems in them, and runs five rules against them. Four of the rules catch a problem and one passes, so you can see both outcomes on the dashboard. The sample lives inside the app, creates nothing in your own databases, and can be removed again from the same page.