Review the FAQs to learn more about API Tasks.

1. What are the different steps involved in creating and using API Tasks in forms (Real-time)?

API Task in a form

2. What are the different steps involved in creating and using API Tasks in process (On-demand)?

API Task in a process

3. How does an API Task execute compared to a Task Bot?

API Tasks execute in the Automation 360 Cloud, unlike Task Bots, which use the bot runner devices on the customer's infrastructure. API Tasks do not perform desktop interactions like Task Bots and can start quickly and potentially run faster than Task Bots.

In addition, the product is General Data Protection Regulation (GDPR) and California Consumer Privacy Act (CCPA) compliant. API Task is also part of the platform that is SOC 2 compliant. All API Task data is encrypted and we also provide encryption for both moving and static data.

4. Is it possible to read or write to a file in API Tasks?
Yes, you can read or write to a file stream within API Tasks. The REST Web Services package now supports reading a FileStream through a variable, allowing you to seamlessly interact with files stored in online storage like OneDrive. While we currently only support OneDrive, our roadmap includes plans to integrate with other popular cloud storage services.
5. Do API Task support custom packages created using the Package SDK?
No, API Task does not currently support packages built using the Package SDK.
6. If one of the devices goes down in my API device pool, would another be spun up automatically?
Yes, everything is managed for you.
7. Can users create a device pool for API Tasks themselves and assign API Tasks to it?
No, you cannot create your own device pool for API Tasks. The Automation Anywhere Cloud platform automatically provisions and manages API Task containers, scaling them dynamically to accommodate fluctuating workloads.
8. What happens if an API Task runs more than the session limit? Does it fail immediately or is it queued?
  • Real-time mode: If an API Task in the real-time mode (attended – called via forms) exceeds the session limit, it will fail immediately. This is because real-time mode tasks are pre-deployed and ready to execute immediately, and there is no queuing mechanism for tasks that cannot be executed due to concurrency constraints.
  • On-demand mode: If an API Task in on-demand mode (unattended – called via processes) exceeds the session limit, it will be queued and processed when a session becomes available. This means that the task will not fail immediately, but will wait until a session is freed up before it can be executed.
9. What is the lifespan of the devices that spin up?
Your API device pool's resources are available for the duration of your configuration.
10. Does anything need to be set up on the deployment machine?
API Tasks execute on the Automation 360 Cloud devices; therefore, no customer devices on their infrastructure need to be set up.
11. Do I need to perform any device installation?
Cloud devices run on clusters managed by Automation Anywhere. There is no device installation required by the customer.
12. Is there a timeout for the API Task execution?
For real-time API Tasks, the timeout is 1 minute. For on-demand API Tasks, the timeout is 15 minutes.
13. What kind of data and for how long does the information persist on the cloud devices?
Transient data does not persist; only the API Task metadata is saved.
14. Why the run option is not available in the API Task editor?
API Tasks cannot be deployed to your local PC. Therefore, the API Task editor does not include a run option. However, you can invoke API Tasks through forms, processes, endpoints, or by using the Run now option.
15. How do I find errors in my API Task?
To verify if your API Task logic is correct, you can use the debugger tool and watch the output variables. You can also use the error handler package actions to check for any errors in the API Task.
16. What is pre-warming?
Pre-warming is the process of pre-installing the API Task logic and commands on Cloud devices preparing it to handle incoming requests. During warming, all API Task dependencies including libraries are installed beforehand. This is beneficial since it eliminates the need to check for dependencies before deploying the API Task. Hence, API Tasks run with very low latency. API Tasks that are invoked in a form must be pre-warmed on the Cloud device for it to execute successfully. Before you execute a process that has the form (API Task attached), you can check if the API Task is pre-warmed in the audit logs. For more details, see View API Task run time audit logs.
17. Can I create a device pool for API Tasks and assign API Tasks to it?

No, you cannot create your own device pool for API Tasks. The Automation Anywhere cloud platform automatically provisions and manages API Task containers, scaling them dynamically to accommodate fluctuating workloads.

18. What happens if an API Task runs more than the session limit? Does it fail immediately or is it queued?
The behavior of API Tasks that exceed the session limit depends on the API Task mode being used:
  • Real-time mode: If an API Task in the real-time mode (attended – called via forms) exceeds the session limit, it will fail immediately. This is because real-time mode tasks are pre-deployed and ready to execute immediately, and there is no queuing mechanism for tasks that cannot be executed due to concurrency constraints.
  • On-demand mode: If an API Task in on-demand mode (unattended – called via processes) exceeds the session limit, it will be queued and processed when a session becomes available. This means that the task will not fail immediately, but will wait until a session is freed up before it can be executed.
19. What are the supported variable types in API Tasks?
The following variable types are supported in API Tasks:
  • String
  • Number
  • Datetime
  • Boolean
  • Credential
  • List
  • Dictionary
  • Record
  • Table
  • Session
  • File
    Note: For more information about packages that support file variables in API Tasks, see File streaming using file variable.
For more information on the variables, see Variables overview.
20. What is the license model for API Tasks?
Every cloud customer receives 100 API Task executions per Control Room as their free base entitlement. The base entitlement also includes two concurrent executions, meaning you can run two API Tasks simultaneously. For higher execution limits and additional concurrency, you can purchase either the Enterprise Platform license or an API Task Overage SKU. For further assistance, contact your customer service manager.
21. How do you view the API Tasks consumption details?
Go to Automation Anywhere Cloud Service section in Administration > Licenses to track the API Task consumption. Click on the Base Executions (Number of API Task executions) or Overage Executions (Number of API Task executions) to view the execution details.
  • The Base Executions lists the consumption details of the free base entitlement.
  • The Overage Executions lists the consumption details that are part of either Enterprise Platform license or an API Task Overage SKU.
22. How does the system calculate the API Task executions?
The system calculates the API Task executions on a per-minute basis. If the execution duration exceeds 1 minute (60 seconds), the execution count increments by one.
For example:
  • An API Task that runs for 30 seconds counts as 1 execution.
  • An API Task that runs for 1 minute and 1 second (61 seconds) counts as 2 executions.
  • An API Task that runs for 3 minutes and 40 seconds (220 seconds) counts as 4 executions.
23. If I include 10 API calls within a single API Task, will it count as one or multiple executions?
It counts as a single execution. You can include any number of API calls within a single API Task.
24. How does the system calculate API Task consumption when an API Task execution is stuck?
If an API Task execution becomes stuck due to any logic or code issues, such as an infinite loop, third-party system non-response, or a response timeout, the system automatically terminates the execution after one minute for real-time API Tasks and 15 minutes for on-demand API Tasks. In these cases, the system counts this as an execution (one execution for real-time API Tasks and 15 executions for on-demand API Tasks) and updates the consumption details on the License page.
If the execution request remains queued due to device unavailability, the system does not count the execution until a device becomes available and the execution is completed.
Note: You can abort the stuck API Tasks if you have the Manage my activity permission. For further assistance, you can contact their customer service manager.
25. When deploying an API Task on-demand within a process, you encounter the error "UNKNOWN: Unable to find the Bot Agent auto registration token." How can you resolve this?
The above error indicates that a registration key is not available for on-demand deployment of the API Task. To resolve this:
  1. Go to Administration > Settings > Bot agent bulk install.
  2. Click Edit, the Generate a new key option is selected by default.
  3. Click Save changes to generate a registration key.