When you have a polling job running on a connector instance for one or more resources and the API calls for that instance have failed a number of times consecutively, this polling job will be automatically disabled.
Some reasons for an API call failing could be:
- User reaching the vendor API limits.
- User modifying the credentials provided while provisioning an instance for the connector.
- Vendor services being unresponsive due to vendor downtime or other vendor issues.
- Token refresh failing.
Continuing to run a poller job when the instance is failing to make successful API calls is redundant as you would not be receiving any successful events. And therefore:
- Our platform automatically disables your polling job for the connector instance if consecutive polls have failed X number of times. The value for X will vary depending on the environment you are on:
- Snapshot - 10
- Staging - 50
- Production - 100
- You will be notified via email when your polling job is disabled. You can take necessary actions based on the information provided in the email.
- You will also see a message that indicates that your polling job is disabled on the edit instance page.
- A list of all the connector instances that have polling jobs disabled by the platform due to multiple consecutive unsuccessful API calls can be fetched via an API
GET/instances/polling-disabled. This helps a user understand which of the instances in their account have their poller jobs disabled. - You can get the data relevant to any jobs that have been running after polling jobs disabled by the platform using the API
GET/jobs/polling-disabled.
Note: Your polling job may be running on an connector instance for multiple resources. However, the polling job gets disabled by the platform even if any one of the resources reach the error count threshold for that particular environment.
You can take appropriate actions based on the error message to start making successful API calls again. You can:
- Check if you have entered the right credentials while provisioning your connector instance.
- Check if your sandbox is working fine.
- Re-authenticate to see if vendor API calls are working as expected.
- Or Use the API
PUT/jobs/{id}/enableto resume poller jobs that have been disabled by providing the jobID.
