The ngrok Traffic Inspector provides observability for traffic towards APIs or services to better understand what is happening and help identify any issues. Since it was previewed earlier this year, the Traffic Inspector has acquired new capabilities based on user feedback and is now officially available through the ngrok dashboard.
Traffic Inspector provides a view of the HTTP traffic across the endpoints belonging to an ngrok account. The inspector captures all metadata associated with requests by default and can be configured to capture full request and response bodies. Additionally, it can replay events to reproduce anomalous behavior.
As ngrok product manager Russ Savage explains, Traffic Inspector was created to provide observability to several ngrok agents, including SSH Reverse Tunnel, ngrok SDKs, and Kubernetes Operator. Since those agents did not provide observability on their own, ngrok created the Traffic Inspector as a gateway that collects traffic from all endpoints. This approach has the added benefit of consolidating all observed data into a single location.
During the three-month developer preview, ngrok has implemented a few new features in its Traffic Inspector, including live updating, additional retention days, replay requests with modifications, and filtering by specific time range.
Live update is useful to debug applications in real-time and pauses automatically when you filter events by date to investigate production issues. Ngrok also extended the traffic retention window from three days to a maximum of 90. As Savage explains, the initial limitation to three days during the preview phase was meant to ensure the system kept its performance even for users with a significant amount of traffic while giving ngrok the chance to analyze and optimize data storage and queries.
Another useful new feature is the possibility of replaying requests with modifications. Replaying requests directly from the ngrok dashboard makes it easy to reproduce events without going through possibly complex activation scenarios, e.g., accessing a third-party service to trigger a webhook in your API, or waiting for an anomalous event to happen again.
Being able to modify the requests before replaying them adds another level of control to investigate how a service behaves in a certain way. This includes changing headers, updating credentials, or changing values in the body, as well as sending the request to a different endpoint with a potential fix.
As a final note, ngrok is already at work on new features for its Traffic Inspector, including advanced filtering and better integration with cloud edges and endpoints.