Quick question about the integration behavior between Architect Data Actions and the Performance Dashboard in our EU-West BYOC environment.
- Environment: Genesys Cloud EU-West BYOC
- Architect Flow Version: 4.2.1
- Data Action Type: External REST API Call (POST)
- Dashboard View: Real-time Queue Activity & Agent Performance
- Issue: Intermittent HTTP 504 Gateway Time-out on Data Actions
Our team is experiencing a discrepancy in metric reporting when Architect flows execute external Data Actions that exceed the standard timeout threshold. The specific issue involves a Data Action designed to fetch customer context from an external CRM system. When the external endpoint responds slowly (between 10-15 seconds), the Data Action fails with a 504 error.
The core concern is not the failure of the individual interaction, but the subsequent impact on the Performance Dashboard. It appears that conversations involving failed Data Actions are not being correctly categorized in the real-time queue activity metrics. Specifically, the ‘Average Handle Time’ (AHT) calculation seems to exclude these failed interactions prematurely, leading to an artificially low AHT during periods of high external system latency. This creates a significant gap between the operational reality reported by agents and the metrics visible to management in the Performance Dashboard.
We have verified that the Data Action configuration includes appropriate retry logic, but the dashboard metrics do not reflect the retry attempts or the eventual success/failure state accurately. The conversation detail view shows the Data Action failure, yet the aggregated performance views remain inconsistent.
Has anyone encountered similar metric alignment issues when using Data Actions in complex Architect flows? We are looking for best practices to ensure that dashboard metrics accurately reflect the full lifecycle of interactions involving external integrations, particularly in a BYOC environment where network latency can be variable. Any insights into configuring Data Actions to minimize this reporting discrepancy would be highly appreciated.