> ## Documentation Index
> Fetch the complete documentation index at: https://docs.clustr-ai.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Bulk jobs

> How large an import can be, and how your jobs run side by side

## Job size

One import job takes up to 1,000 LinkedIn URLs or connectors. A longer list answers `400`:
send it as several requests.

Your contract can set a lower largest job, and a monthly total. A job over either is
refused with `422` and code `ORG_LIMIT_EXCEEDED`. See [Usage and limits](https://docs.clustr-ai.com/enterprise/usage-and-limits.md).

A person listed twice in one request is processed and counted once.

## Retry safely

Add an `Idempotency-Key` header to an import. If your call times out, send it again with
the same key. You get the first job back, with the header `Idempotent-Replayed: true`,
and no second job starts. The key is scoped to the workspace, and can be up to 200
characters.

```bash theme={null}
curl -X POST https://api.clustr-ai.com/api/public/network/import/connectors \
  -H "Authorization: Bearer $CLUSTR_ORG_KEY" \
  -H "X-Clustr-Tenant: client-1234" \
  -H "Idempotency-Key: client-1234-batch-001" \
  -H "Content-Type: application/json" \
  -d '{ "linkedin_urls": ["https://www.linkedin.com/in/example"] }'
```

## When several jobs run at once

Each client's jobs keep moving while another client runs a bulk job, and when only one client
has work, its jobs run at full speed. A long list takes longer to finish than a short one. If
a job seems slow, ask your Clustr contact.

## Follow a job

A job can be `QUEUED`, `PROCESSING`, `COMPLETED`, `FAILED` or `CANCELLED`. Read it with
[Get Job Status](https://docs.clustr-ai.com/api-reference/get-job-status.md), list a workspace's jobs with
[List Jobs](https://docs.clustr-ai.com/api-reference/list-jobs.md), or have the organization
[webhook](https://docs.clustr-ai.com/enterprise/webhooks.md) call you with `job.completed` and `job.failed`.
[Cancel Job](https://docs.clustr-ai.com/api-reference/cancel-job.md) stops a job and keeps everything it already imported.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.