The Genesys Cloud data actions integration provides static actions and also allows you to create custom actions that use the Platform API. Custom actions interact with data from the Platform API through the request and response schemas that you define. Genesys Cloud defines limit thresholds for data actions usage. The limits, usage of data actions, and performance characteristics of the remote endpoints that the data actions connect with are related. 

The Data Actions Performance Summary view shows some of the metrics that outline the function of remote endpoints. Some statistics relate to the following limits that are defined for the data actions:

• actions.concurrent.executions.max
• actions.integration.max
• actions.total.max
• requests.volume.max

For more information, see Limits in the Genesys Cloud Developer Center.

You can interpret the following statistical details of the data action from the Data Actions Performance Summary view based on the limits:

• 95% Duration and 99% Duration: This data gives an insight into information about any delay in response from a remote endpoint. The statistics relate to the actions.concurrent.executions.max limit. The data action that runs longer consumes more concurrency. If you experience Concurrency Rate Limit Errors, check the duration statistics to confirm whether the errors are due to latency in the remote endpoint to which the data action connects.
• Volume Rate Limit Errors: The number of data actions for which the execution was rejected because the organization is above the allotted volume limit. 
• Concurrency Rate Limit Errors: Concurrency rate limit is a restriction that limits the number of simultaneous requests that can be made to an API. For example, if the  rate limit is set to a maximum of 50 executions, the error occurs when all the 50 slots allotted for the concurrent execution of a data action are consumed. Your volume of requests per minute can be considerably moderate or high, but the duration taken for each data action response can consume the allotted slots faster and result in the concurrency rate limit error.
• Remote Endpoint Errors: This error occurs when you get a non 200 level response from the remote endpoint. For example, when the data action finds no search results or more than one search result, the expected behavior is a 404 response. You can filter such failed 404 responses.
• Processing Errors: The data action executions fail because of the internal processing steps. The steps could include responses that don’t conform to the stated output schema. To investigate further on this error, check out the results of the test operation of your data action.

The Data Actions Performance Summary view and Data Actions Performance Details view help you gather information about your data actions performance, detect and diagnose errors, understand the causes of failures, and analyze why certain actions are not working as expected.