Your first connection and check¶
A walkthrough from a fresh account to your first passing (or failing) check result. Screen-by-screen guides for each part of the app are coming in a later revision of this site — this page is the fast path.
1. Sign in¶
Sign in at your CatalystData URL with the account details from your Marketplace subscription (see Getting started if you haven't subscribed yet).
2. Add a connection¶
Go to Connections → New Connection. You'll need:
- A connector type — see Connect a data source for the currently supported connectors and what each one needs
- Host, database, and an auth type appropriate to that connector
Click Test Connection before saving to confirm CatalystData can reach the data source.
3. Scan the connection¶
Go to Scans → New Scan, pick the connection, and run it. A scan discovers the connection's tables and columns — you'll use this to pick what to check, and CatalystData uses it to suggest checks for you.
4. Review suggested checks¶
After a scan completes, go to Connections, open the connection, and
look at its Recommendations. CatalystData suggests checks based on what
the scan found — a null check on a mostly-non-null column, a unique
check on what looks like a primary key, and so on. Accept the ones that make
sense; each becomes a check you can edit later.
5. Add a check by hand¶
Go to Checks → New Check, choose table-level or column-level, pick a
table (and column, for column-level checks), and a check type. See the
check reference for every check type and its
parameters. A good first check: row_count on your most important table,
with a min_count that would catch an empty load.
6. Run it¶
Checks run automatically on their schedule, or click Run Now to run one immediately. Go to Results to see the outcome — pass, fail, error, or warning — along with the metric value and threshold that produced it.
7. Get notified when it fails¶
Go to Alerts → Channels → Add Channel and configure email, a webhook, or a Slack/Teams incoming webhook. Then go to Alerts → Rules → Add Rule, link the channel to your check (or to all checks in a schedule), and choose a condition — fire on any failure, only on errors, or after a set number of consecutive failures. See Alerts and scheduling for the full set of channel types and conditions.
Next¶
- Check reference — every check type and what it catches
- Data contracts — bundle a schema and a set of checks into a versioned contract you can validate a table against
- Roles and permissions — invite your team with the right access level