The main purpose of this feature is to provide a simplified and efficient way to investigate errors that occur within the Eddie system, whether in flows or in specific conversations. We currently face significant limitations in obtaining data when it comes to Eddie's errors, making the debugging process a substantial challenge. Not only errors, even in performance (e.g., where a ticket stopped in the flow), it is difficult for us to carry out analyses and suggestions to help the customer (or even to build products based on that data).

Faced with this scenario, the proposed evolution of the feature seeks to implement a robust system for tracking errors and logs associated with Eddie. This data will be carefully stored in a database (available in Lake and Metabase), allowing for more detailed and comprehensive analyses to be carried out both by the developers and other teams involved.

The focus will not only be limited to the error data itself, but will also include correlative information, such as user conversion metrics within operational flows, ticket analysis, and others.

Thus, by consolidating and expanding visibility of this information, it is expected not only to improve problem solving capacity, but also to optimize processes and enrich the understanding of system behavior. The initiative thus aims to transform a context of ignorance into an opportunity for more informed and strategic decisions.