Live limit reached

Your organization is at one of its live limits, so the request was refused. Nothing was created. The limit frees as soon as one of your running resources ends.

https://docs.asteroid.ai/errors/limit-reached

This URL is the type of the error. Match on type, not on message. The message text can change.

When you get it

Request	`capability`
Start an execution	`concurrent_executions`
Start an environment	`concurrent_environments`

A resource counts while it runs. It stops counting when it ends, or when it passes its deadline:

An execution counts until its workflow timeout, plus 5 minutes.
An environment counts until its expiry time.

Response

The status is 429 Too Many Requests. The body follows RFC 9457.

{
  "type": "https://docs.asteroid.ai/errors/limit-reached",
  "title": "Live limit reached",
  "status": 429,
  "detail": "Your organisation is at its limit of 5. Retry once one ends",
  "instance": "/agents/v2/workflows/0b6e.../execute",
  "capability": "concurrent_executions",
  "used": 5,
  "limit": 5,
  "code": 429,
  "message": "Your organisation is at its limit of 5. Retry once one ends"
}

string

required

Always https://docs.asteroid.ai/errors/limit-reached for this problem.

string

required

Live limit reached.

integer

required

429.

string

required

A sentence you can show to a person.

string

The request path that was refused.

string

required

The limit you reached. See When you get it.

integer

required

Resources of this kind running when the request arrived.

integer

required

Your organization’s limit for capability.

integer

The same as status. Kept for older clients.

string

The same text as detail. Kept for older clients.

Handle it

Queue the execution instead

For executions, send "onCapacityLimit": "queue". The execution waits as queued and starts by itself, oldest first, when a slot frees. You never see this error. See Queue when capacity is full.

Or retry later

Wait for a running execution or environment to end, then send the request again. Back off between attempts.

Raise the limit if you hit it often

Limits come from your plan. Contact support to change them.

const LIMIT_REACHED = 'https://docs.asteroid.ai/errors/limit-reached';

type LimitReached = { type: string; capability: string; used: number; limit: number };

const { data, error, response } = await workflowExecutePost({
  path: { workflowId },
  body: { inputs },
});

if (response.status === 429 && error?.type === LIMIT_REACHED) {
  // capability, used and limit say which limit is full.
  const problem = error as unknown as LimitReached;
  console.log(`${problem.capability}: ${problem.used} of ${problem.limit} in use`);
}

import json

from asteroid_odyssey.exceptions import ApiException

LIMIT_REACHED = "https://docs.asteroid.ai/errors/limit-reached"

try:
    workflows_api.workflow_execute_post(
        workflow_id=workflow_id,
        workflows_workflow_execute_request=request,
    )
except ApiException as e:
    problem = json.loads(e.body or "{}")
    if e.status == 429 and problem.get("type") == LIMIT_REACHED:
        # problem["capability"], problem["used"] and problem["limit"] say which limit is full.
        print(f'{problem["capability"]}: {problem["used"]} of {problem["limit"]} in use')
    else:
        raise

Response	Meaning
`403` with `capability_required`	Your plan has a limit of `0` for this resource. Retrying does not help.
`429` with any other `type`	Too many requests arrived at once to check the limit, or the execution queue is full. Retry after a short wait.

Get Started

Errors

Reference

Webhooks

When you get it

Response

Handle it

​When you get it

​Response

​Handle it

​Related errors

When you get it

Response

Handle it

Related errors