Skip to main content
Syncs are the background runs that keep your knowledge stores current. You monitor them because stale data means wrong agent answers — if a sync fails silently, your agent keeps serving yesterday’s content while the source has moved on. This page gives you environment-wide visibility into every ingestion run. A sync is a single ingestion run for a data source. The syncs endpoints give you an environment-wide view of ingestion health: what ran, what succeeded, and what needs attention.

​
Sync states

​
List all syncs (environment-wide)

​
Response

​
array
Array of sync objects.
​
string
Unique sync identifier.
​
string
The data source this sync belongs to.
​
string
One of pending, running, completed, failed.
​
string
ISO 8601 timestamp when the sync started (null if pending).
​
string
ISO 8601 timestamp when the sync finished (null if not terminal).
​
integer
Current page number.
​
integer
Items per page.

​
List syncs for a data source

​
Errors

​
Production recommendations

  • Review the environment-wide sync view on a schedule so stale or failing sources don’t go unnoticed.
  • Set each source’s sync cadence to match how often the underlying data changes.
  • Treat repeated failed syncs as an alert; pause and reconfigure the source rather than letting it retry indefinitely.

​
Rate limits

10,000 requests per organization per hour. Pagination: page >= 1, page_size 1–100 (default 20).
Review the environment-wide syncs view regularly. A completed status confirms ingestion finished, but check item counts to verify completeness.
When a sync fails, inspect the error details in the sync object. Common causes:
  • Bucket permissions changed (missing s3:GetObject / s3:ListBucket).
  • Objects exceed max_object_size_mb.
  • Unsupported file formats in the bucket.
Fix the underlying issue, then trigger a new sync from the Connectors endpoint.

Connectors

Configure the sources that sync.

Knowledge Stores

Where synced content lands.

Documents & Uploads

Add files directly.

Data Platform Overview

The full data model.

​
Next steps

Continue to Context Bundles.