Does anyone know the specific payload size thresholds or architectural limitations that trigger a 504 Gateway Timeout when executing custom data actions within a multi-tenant AppFoundry integration? We are deploying a premium application that processes high-volume customer interaction logs, and the execution pipeline is failing consistently when the input JSON array exceeds approximately 500 records. The data action is designed to enrich these records via a downstream REST API call to our partner platform, but the Genesys Cloud Platform API returns a 504 error before the action completes its processing cycle. The environment is configured with standard OAuth 2.0 client credentials for multi-tenant access, and the data action is invoked via an Architect flow that batches these records every 10 seconds. We have verified that the downstream endpoint is responsive and capable of handling the load, as confirmed by direct curl tests from our staging environment. The issue appears to be strictly within the Genesys Cloud execution context, suggesting a timeout limit on the data action worker thread or a restriction on the maximum execution time for serverless functions within the AppFoundry ecosystem. We are using the latest version of the Genesys Cloud Python SDK for our backend logic, and the data action schema is validated against the OpenAPI specification without any structural errors. The error logs in the Genesys Cloud admin console do not provide detailed stack traces, only indicating that the request timed out waiting for a response from the data action service. This behavior is blocking our ability to scale the integration to handle peak call volumes, as batching smaller payloads introduces significant latency in the overall workflow. We need to understand if there is a configurable timeout setting for data actions or if we must restructure the architecture to use asynchronous webhooks instead of synchronous data action calls. Any insights into the internal execution limits or best practices for handling large datasets in this context would be highly appreciated.