Platform Defaults FAQ
Answers to common questions about Platform Defaults in Qualytics.
Access and Roles
Who can edit Platform Defaults?
Only users with the Admin user role. The Defaults tab is hidden from Managers and Members in the Settings tab bar. Users with the Manager role can call GET /api/settings/global-defaults to inspect the current values via the API, but they cannot save changes.
Do team permissions affect Platform Defaults?
No. Platform Defaults are workspace-wide, so access is decided by user role only. Editor, Author, Drafter, Viewer, and Reporter team permissions have no effect on this tab.
Behavior
What happens when I save a Platform Default?
The new value becomes the fallback at the bottom of the precedence chain. Every new datastore created after the save inherits it, and every new operation on a datastore that has no explicit override picks it up. Existing datastores and containers keep their current values unless you also turn on the Apply to existing datastores toggle at save time.
What does the Apply to existing datastores toggle do?
On the Observability tab this toggle reads Apply to existing datastores and containers; on Profile and Scan it reads Apply to existing datastores. When on, saving the tab pushes the new defaults out to every existing source datastore. For the four Observability defaults, it also pushes them to every non-deleted container. This is a hard overwrite: values that individual authors previously set on their own datastores or containers are wiped and replaced with the new defaults. The toggle is off by default and resets to off after each save.
Why did my per-datastore setting change without me touching it?
An Admin most likely saved a Platform Default with the Apply to existing datastores toggle on. That propagation overwrites the corresponding values on every existing datastore for the settings in the group that was saved. If you are not sure when the change happened, ask your Admin to confirm.
Observability
Why does turning on the AI check control also turn on Tracking?
Tracking and automation are coupled. AI-generated observability checks depend on the daily metrics that Tracking produces, so Qualytics enables Tracking automatically whenever you enable an AI check control. The same rule applies in reverse: turning Tracking off disables the AI check control, since there is no data to base the check on.
Can I keep AI checks off but still track volume or freshness?
Yes. Leave the Tracking toggle on and untoggle the corresponding Create AI ... checks control. Tracking will keep recording daily metrics, and no AI check will be created automatically.
Profile
How is AI Effort resolved at profile time?
The precedence chain for profile operations is: value set on the operation itself, then the datastore-level default, then the Platform Default. Qualytics ships a medium global default that every new datastore inherits at creation.
What's the difference between activating AI checks automatically and keeping them as drafts?
When Activate AI managed checks is on, AI-generated checks are enabled as soon as the profile completes and start running against your data on the next scan. When it is off, the checks are saved as drafts. Drafts are visible in the platform but must be reviewed and activated before they participate in scans.
Scan
What is the record anomaly rollup?
When a discrete check has more failing records than the Maximum Record Anomalies per Check threshold, Qualytics emits a single rolled-up anomaly for the check instead of creating one record anomaly per failing row. This keeps the anomaly list manageable when a systemic problem produces thousands of individual failures. When the number of failing records is at or below the threshold, each failing row stays as its own record anomaly.
What are source examples and why are they capped?
Source examples are the individual source records saved to the enrichment datastore for each failing check, so you can see representative data for the anomaly. The Maximum Source Examples per Anomaly setting caps how many examples are saved, which controls the cost of storing enrichment data. When the number of failing records is below the cap, the count of saved examples equals the number of failing records; at or above the cap, only the cap number is saved.
API
Is there a way to see what changed after a save?
The response body of a successful PATCH /api/settings/global-defaults request contains a settings object with the full current defaults after the save. Compare it against the values you had before the request to see what changed. When propagate=true is used, the response also includes a propagation object with datastores_updated and containers_updated counts.
Can I set defaults per team or per datastore group?
No. Platform Defaults are workspace-wide. To scope different defaults to different datastores, set the per-datastore values on those datastores directly. Datastore-level values always win over Platform Defaults in the precedence chain.