Rate limits and concurrency

Launch limits

By default, launch limits apply per learner within your organization over a rolling five-minute window:

Membership	Launches per window
`free`	1
`premium`	15

These limits are configurable. The API response gives the wait time for your request. Ending a lab does not reset the launch window. Membership comes from your server access policy. The API returns a rate-limit error with a wait duration. The SDK exposes that duration as retryAfter in seconds.

Organization capacity

An organization also has a limit on active deployments. The default cap is 75. This can be changed for your organization. Contact us for help. The SDK returns concurrency_limited with activeCount and limit. Teardown must release capacity before another launch can use that slot. A successful end request does not promise immediate slot availability.

Other operations

The launch window does not count status reads, flag submissions, or guide answers. Discussion reads and writes have separate limits. A five-second status interval is the SDK default. Repeated launch requests do not make provisioning faster.

Handle a limit

import { CybrLabsError } from '@cybr/labs-sdk'

try {
  await cybr.deployments.launch({ labId, learnerId })
} catch (error) {
  if (error instanceof CybrLabsError && error.code === 'rate_limited') {
    console.log('Wait seconds:', error.retryAfter)
  }
  throw error
}

The UI can show a wait and let the learner request another launch. A long automatic wait is not required. The SDK only retries short rate-limit waits within its retry configuration. See client configuration. For a different organization cap, contact support@cybr.com.

SDK setup

SDK guides

SDK reference

Rate limits and concurrency

Launch limits

Organization capacity

Other operations

Handle a limit

​Launch limits

​Organization capacity

​Other operations

​Handle a limit

Launch limits

Organization capacity

Other operations

Handle a limit