Manage Knowledge Auto-Sync

What Auto-Sync does

Auto-Sync watches files added through Documents, Applications, users, or Flow Builder and processes matching files automatically for Knowledge search.

An Auto-Sync rule answers four questions:

1. Which files should match?
2. Which file types are allowed?
3. Which processing preset should be used?
4. When should processing start?

Open the page from Knowledge Documents → More → Auto-Sync Rules. Use Back to Knowledge to return to /knowledge-documents and Refresh to reload the active tab.

Understand the four tabs

The New Rule button is shown only on the Rules tab.

Before creating a rule

• Create at least one processing preset from the Processing Preset menu in Knowledge Documents.
• Decide whether the rule should process only new matching files or also existing files through Run Now.
• Identify the narrowest useful scope.
• If scopes can overlap, keep Skip files that have already been processed enabled.

Create a sync rule

Open Rules, click New Rule, and complete the six-step wizard.

Step 1: Identity

Configure:

• Rule Name — required;
• Description — optional context for other administrators;
• Enabled — whether the rule can create new automatic jobs after it is saved; and
• Priority — the rule's ordering value.

Use a name that states both the source and purpose, such as HR policies to Knowledge. You can save a disabled rule, test its matching behavior, and enable it later.

Step 2: Scope

Choose where matching files come from:

For App / Page / Field, leaving Application, Page, or Field empty keeps the match broader at that level. Use the breadcrumb-style Scope summary to verify the final selection.

For User / Role, the page lets you combine selected roles and users. Clear a user chip or click a selected role again to remove it.

For Flow Builder Workflow, leave Node Label empty to match all upload nodes in the workflow. Enter a label only when the rule must target one upload step.

Step 3: File Filters

Select any allowed extensions:

pdf, docx, doc, md, xlsx, csv, pptx, txt, and html.

Leave all extensions unselected to allow every supported type.

When pdf is selected, PDF Page Limit appears. Set Max Pages to 0 for no limit. A positive value skips PDFs that exceed that page count.

Step 4: Pipeline

Select one required Target Preset. The selected preset defines parsing, chunking, enrichment, and embedding behavior for matching files.

The current wizard accepts one target preset per rule. If no preset is available, create one from the Processing Preset menu in Knowledge Documents, then return and refresh the page.

Step 5: Trigger

Choose when matching files are processed:

• Immediately — starts when the file is uploaded;
• After a delay — waits a selected number of minutes or hours; or
• On a schedule — processes at a recurring hourly, daily, or weekly time.

For a schedule, configure the repeat interval, day when weekly, time when applicable, and Zone. Review the human-readable schedule summary before continuing.

Under Reliability, configure:

• Retry attempts if it fails;
• Wait between retries (minutes); and
• Skip files that have already been processed (recommended).

The last option compares file content and helps prevent duplicate processing when uploads or rules overlap.

Step 6: Review

The Review step summarizes:

• rule name;
• enabled or disabled status;
• matching scope;
• trigger timing;
• target preset;
• priority; and
• duplicate handling.

Click Create Rule. A rule name and target preset are required. When editing an existing rule, the final action is Save Changes.

Read and manage rule cards

Each card shows the rule name, description when present, trigger, scope, priority, and target preset. Use the controls to:

• enable or disable the rule;
• Dry-run;
• Run Now (backfill);
• Edit; or
• Delete.

Disabling a rule stops it from creating new automatic jobs. It does not cancel work that has already started.

Deleting a rule also removes all jobs associated with that rule. This is different from disabling it and requires confirmation.

Test a rule with Dry-run

Click Dry-run to evaluate the rule against existing Documents content without processing or changing anything.

The dialog shows:

• the number of matching documents; and
• a sample list with filename, content type, and size when available.

If the count or sample is wrong, close the dialog and narrow the rule's scope or file filters before enabling it.

Process existing files with Run Now

Click Run Now (backfill) to queue matching files that already exist. The confirmation reports both the number of matching documents and the number of queued jobs.

Use Job History to monitor the result. Normal automatic triggering applies to later uploads; Run Now is the explicit action for existing content.

Monitor Job History

Each job row can show:

• readable status;
• filename or document number;
• target preset identifier;
• number of indexed items;
• duration;
• creation time; and
• a readable error when processing failed.

Current statuses include Succeeded, Failed, Running, Scheduled, Pending, Cancelled, and Skipped (duplicate).

Available actions depend on status:

• click Retry for a Failed job;
• click Cancel for a Scheduled or Pending job; and
• wait or refresh for a running job.

If the tab says No jobs have run yet, enable a rule and upload a matching file, or use Run Now for existing files.

Read Stats

Stats covers the last 24 hours and can show:

• a count for each job status;
• Avg pipeline duration (succeeded); and
• Top Rules (last 24h) with each rule's job count.

The average includes successful jobs, not every attempted job. When there has been no activity during the window, the page shows an empty-state message instead of zero-filled cards.

Govern presets

Open Presets to manage presets used by documents and sync rules.

Set the default preset

The Document processing preset panel identifies the global default used when processing starts without an explicitly selected preset.

• click the star beside a preset to Set as default;
• click the filled star or Clear default preset to remove the default; and
• look for the Default badge to confirm the current choice.

Rename or delete a preset

Use Rename preset to change its display name. Use Delete preset only after reviewing documents and sync rules that may use it; deletion cannot be undone.

This tab does not create presets. Create one from the Processing Preset menu in Knowledge Documents.

Recommended setup pattern

1. Create and test a processing preset on a representative document.
2. Create a narrowly scoped, disabled Auto-Sync rule.
3. Run Dry-run and review its sample.
4. Correct the scope or filters if needed.
5. Enable the rule.
6. Use Run Now only if existing matching files should be included.
7. Review Job History and Stats after initial processing.

Troubleshooting