Pagination

List methods in the Harbor SDK return a page of results plus a cursor for the next page. All list endpoints use the same cursor pattern.

Basic usage

import { Harbor } from '@harbor/sdk';

const harbor = new Harbor({ secretKey: process.env.HARBOR_SECRET_KEY });

const page = await harbor.events.list({
  workspaceId: 'ws_018f3a2e4b9c',
  limit: 25,
});

for (const event of page.data) {
  console.log(event.id);
}

if (page.hasMore) {
  const next = await harbor.events.list({
    workspaceId: 'ws_018f3a2e4b9c',
    limit: 25,
    startingAfter: page.endCursor,
  });
}

Auto-pagination helper

For batch jobs, use the async iterator so you do not manage cursors manually:

for await (const event of harbor.events.listAuto({
  workspaceId: 'ws_018f3a2e4b9c',
  limit: 100,
})) {
  await archiveEvent(event);
}

The helper respects rate limits and retries transient failures between pages.

Response shape

Field	Description
data	Array of resources for this page
hasMore	true when another page exists
endCursor	Pass as startingAfter on the next request

Default limit is 20. Maximum is 100.

Sort order

List endpoints return newest resources first unless documented otherwise. You cannot sort by arbitrary fields through the public API.

tip

When syncing to a data warehouse, persist endCursor after each successful page so a failed job can resume without re-fetching earlier rows.

Common mistakes

• Reusing an expired cursor. Cursors expire after 24 hours. Restart from the beginning or from a saved checkpoint timestamp.
• Setting limit above 100. Requests above the maximum return invalid_request.
• Parallel page fetches. Harbor list endpoints are cursor-sequential. Fetch one page at a time.

Next steps

See Events (SDK reference) for list options. Watch Rate limits when iterating large datasets.