How Platform Defaults Work¶
This page covers the anatomy of the four default groups, the precedence chain that resolves a setting at operation time, cascade rules that couple related settings, and how the optional propagation to existing datastores behaves.
The Four Default Groups¶
The Defaults tab groups every workspace-wide default into four sections, in the order the left rail shows them: Observability, Sync, Profile, and Scan. Each section has its own Save button; changes on one section do not affect the others until you save the corresponding section.
Observability Defaults¶
Controls the defaults for daily tracking of container metrics and whether AI-generated observability checks are created when tracking is on.
| Setting | Control | Qualytics default | Description |
|---|---|---|---|
| Volume Tracking | Toggle | Off | Turns on daily recording of volume metrics for each table or file. |
| Create AI volumetric checks | Checkbox nested under Volume Tracking | Off | When on, Qualytics automatically creates an AI-generated volumetric check for each tracked container. Disabled when Volume Tracking is off. |
| Freshness Tracking | Toggle | Off | Turns on daily recording of the last time data was added or updated for each container. |
| Create AI freshness checks | Checkbox nested under Freshness Tracking | Off | When on, Qualytics automatically creates an AI-generated freshness check for each tracked container. Disabled when Freshness Tracking is off. |
Cascade rule for tracking and automation
Tracking and automation are coupled. Turning Volume Tracking off automatically turns Create AI volumetric checks off. Turning Create AI volumetric checks on automatically turns Volume Tracking on. The same coupling applies to the freshness pair. This coupling is enforced at save time.
Sync Defaults¶
Controls how many tables or files a Sync analyzes at the same time. It is the one group that holds a single value for the whole deployment rather than a default that datastores and operations can override.
| Setting | Control | Range | Qualytics default | Description |
|---|---|---|---|---|
| Tables or Files Analyzed at Once | Numeric input with a preset menu (4, 16, 32, 64, Custom) | 1 to 256 | 16 | How many tables or files Qualytics analyzes in parallel while syncing a datastore. A higher number can make a sync of a datastore with thousands of tables finish sooner, but it puts more load on the source system. |
The value is a whole number. Typing one outside the range shows the hint Value must be between 1 and 256, and the field clamps to the nearest bound when it loses focus, so an out-of-range value is never saved.
Three things set this group apart from the other three:
- It has no Apply to existing datastores toggle. There is nothing beneath it to propagate to: the limit is one shared allowance across every Sync in the deployment, not a per-datastore setting. The side card shows a single Global level for it.
- Saving it takes effect on the next Sync, with no restart. Each Sync carries the current value with it, and the platform resizes its allowance as soon as a Sync that carries a different value starts.
- It belongs to no datastore. The scoping rule for the other groups, source datastores versus enrichment datastores, does not apply. Every Sync anywhere in the deployment shares the same limit.
Note
The Sync entry appears in the left rail only once your deployment's platform serves this setting. On a deployment that does not yet, the rail shows the other three groups and nothing is hidden behind an empty panel.
Profile Defaults¶
Controls how Qualytics AI generates checks during a profile operation.
| Setting | Control | Qualytics default | Description |
|---|---|---|---|
| Effort | Slider with six named stops | Medium | How much effort AI should put into generating checks. Higher effort considers broader categories of patterns, applies deeper analysis, and takes longer to complete. See the effort table below. |
| Activate AI managed checks | Checkbox under an Oversight section | On | When on, AI-generated checks activate as soon as the profile completes. When off, the checks are saved as drafts and must be reviewed before they take effect. Disabled when Effort is Off. |
The six Effort stops, from lowest to highest:
| Effort | Description |
|---|---|
| Off | No AI checks will be managed. Qualytics still profiles your data, but will not create or update checks automatically. |
| Low | Fastest. Catches basics like missing values, negative numbers, future dates, and unexpected values in known-option fields. |
| Medium | Balanced. Adds checks for unusual value ranges, unexpected text formats, and fields that should be unique. |
| High | Deeper. Adds checks that spot mismatches between related fields. |
| Extra High | Intensive. Adds predictive checks that flag values inconsistent with others in the same row. |
| Max | Maximum. Adds fine-grained checks for patterns, completeness, and date precision. |
Each stop includes everything from the previous stop.
Scan Defaults¶
Controls anomaly rollup and the number of source records written to the enrichment destination during a scan.
| Setting | Control | Range | Qualytics default | Description |
|---|---|---|---|---|
| Maximum Record Anomalies per Check | Numeric input with preset chips (1, 10, 100, 1K) | 1 to 1000 | 10 | How many record anomalies can be created per check before they are grouped into one rolled-up anomaly. |
| Maximum Source Examples per Anomaly | Numeric input with preset chips (1, 10, 100, 1K, 1B) | 1 to 1,000,000,000 | 10 | How many source records are stored in your enrichment destination as examples when a check fails. |
Enrichment Datastores¶
Observability and Profile defaults apply only to source datastores, not enrichment datastores. Scan defaults apply to source and enrichment datastores. Enrichment datastores start with observability tracking disabled and AI Effort Off; you can configure their own Profile settings separately. The automatic Profile after an enrichment write uses a lightweight configuration with AI check generation off, independently of platform Profile defaults.
The Precedence Chain¶
Every setting resolves through a chain from most-specific to least-specific. The first level in the chain that has a value wins; if no level sets a value, Qualytics' shipped default is used.
The chain differs by group:
| Group | Precedence (highest to lowest) |
|---|---|
| Observability | Container > Source Datastore > Global (Platform Default) |
| Sync | Global (Platform Default) only. One value for the whole deployment, with no datastore or operation level to override it |
| Profile | Operation > Source Datastore > Global (Platform Default) |
| Scan | Operation > Datastore (Source or Enrichment) > Global (Platform Default) |
When you edit a Platform Default and save, the new value only becomes the fallback at the bottom of the chain. Any explicit value set on a container, operation, or datastore continues to win, unless you also opt in to propagate the new defaults (covered below).
The Defaults tab shows this chain in a side card labeled Order of precedence so you have it in view while editing.
Cascade Behavior¶
Tracking and automation on the Observability group are coupled at save time. If you submit values that would produce an inconsistent state, Qualytics automatically adjusts them:
- If you turn Create AI volumetric checks on while Volume Tracking is off, Volume Tracking is turned on for you.
- If you turn Volume Tracking off while Create AI volumetric checks is on, the AI checks toggle is turned off for you.
- The same rules apply to the freshness pair.
These adjustments happen silently on save. If you want the AI checks toggle to stay off, leave the corresponding Tracking toggle on and untoggle only the AI checks control.
Applying Defaults to Existing Datastores¶
Above the Save button on the Observability, Profile, and Scan tabs there is an Apply to existing datastores toggle (on the Observability tab it reads Apply to existing datastores and containers). The Sync tab has none, since its value lives at the Global level only. When on, saving the tab does two things:
- Records the new Platform Default for future assets.
- Overwrites the corresponding values on existing datastores in scope: source datastores for Observability and Profile, and both source and enrichment datastores for Scan. Observability propagation also updates non-deleted containers in source datastores.
The propagate toggle is off by default and resets to off after every save. Use it when you want a governance change to apply immediately across your workspace; leave it off when you want the change to only affect assets created after the save.
Propagation is a hard overwrite
When the Apply-to-existing toggle is on, propagation replaces per-datastore and per-container customizations for the settings in that group. Values that individual authors explicitly set on their own datastores are wiped and replaced with the new defaults.
What Happens After You Save¶
- The Save button is only enabled when the current tab has unsaved changes or the Apply-to-existing toggle is on.
- Discard reverts the current tab's edits without touching the other tabs.
- On success, an inline confirmation banner reports how many datastores and containers were updated when propagation was on. Example: Observability defaults saved and applied to 3 datastores and 12 containers.
- If save fails, an inline error banner reports the reason. Nothing on the tab is written until save completes.
Reset to Qualytics Defaults¶
Each tab exposes a Reset button next to the section title that loads Qualytics' shipped values back into the form. The button appears only when the current tab has drifted from Qualytics' shipped values, and the reset is not saved until you click Save on that tab.