Same provider, another attempt.

A recoverable failure is resubmitted to the same provider, on the next proxy in your list. Off until you turn it on.

Resilience

An example run, not a record: the default 2 retries, two of the default trigger codes.
Default
Off
Retries when on
2
Most per provider
10
Attempts per task, all providers
30

01Rule

Only failures a new attempt can clear.

  1. 01The attempt failsWith a code on the trigger list.
  2. 02Next proxyEach attempt moves to the next proxy in task.proxy, wrapping around.
  3. 03ResubmitUp to the limit. A request that was wrong, such as an invalid sitekey, ends on the first attempt.
  4. 04Out of retriesAuto Fallback takes over when it is on.

02Config

Turn it on per key or per request.

settings.retry, in POST /createTask

{
  "settings": {
    "retry": {
      "enabled": true,
      "maxRetries": 2,
      "onErrors": [
        "TASK_TIMEOUT",
        "PROXY_BANNED"
      ],
      "providerRules": {
        "TwoCaptcha": {
          "maxRetries": 1
        }
      }
    }
  }
}
Auto Retry fields
FieldRangeDefault
enabledtrue, falsefalse
maxRetries0 to 102
onErrorsAny trigger codeThe 8 below
providerRulesPer provider: maxRetries, onErrorsNone
settings.allowProxyReusetrue, falsefalse

One proxy with retry on would reuse the same IP, so the request is rejected unless you pass more proxies or set allowProxyReuse.

03Triggers

8 codes retry by default.

Default retry triggers

  • CAPTCHA_UNSOLVABLE
  • TASK_TIMEOUT
  • SERVICE_UNAVAILABLE
  • PROVIDER_UNREACHABLE
  • PROVIDER_RATE_LIMITED
  • PROXY_CONNECTION_FAILED
  • PROXY_BANNED
  • ERROR_PROVIDER_ZERO_BALANCE

04FAQ

Questions about Auto Retry.

01

Does retry switch providers?

No. Retry stays on the same provider. Moving to a different provider on failure is Auto Fallback.
02

Are validation errors retried?

No. Only refundable provider failures continue the chain. A bad request such as an invalid sitekey is terminal on the first attempt.