Skip to content
|
search
Kong Konnect

Diagnosing Latency Issues

Custom reporting in Kong Konnect gives you the power to monitor your API data in detail and export that data to a CSV file. Let’s go through an example situation where you could leverage custom reports.

In this scenario, Tal is a platform owner who supervises the implementation and maintenance of the API Gateway infrastructure that governs all APIs at ACME. A few minutes ago, they received an email that stated that the response latency seemed to be very high and and causing a bad customer experience. To better understand the source of the delays, Tal decides to create two reports to compare the average upstream and Kong latency for all traffic to the production servers in the last hour.

Build reports

You can build a custom report by navigating to analytics icon Analytics in the Konnect menu, then Reports. This brings you to a list of all custom reports in the organization. From here, click New Report to get started.

Kong latency

For the Kong latency report, Tal configures the following:

  • Name: Production - Kong Latency (last hour)
  • Chart type: Line chart
  • Date/Time: Last hour
  • Chart type: Line chart
  • Select a metric: Kong Latency (avg)
  • Group by: Service
  • Choose Granularity: Minutely

Then, they add a filter to filter by the individual runtime group:

  • Filter by: Runtime Group
  • Operator: In
  • Value: prod

Production - Kong Latency (last hour)

Figure 1: Line chart showing the latency over the last hour. The line chart shows the Account API briefly spiking while the Payment API doesn’t spike at all.

Upstream latency

For the Upstream latency report, Tal configures the following:

  • Name: Production - Upstream Latency (last hour)
  • Chart type: Line chart
  • Date/Time: Last Hour
  • Chart type: Line chart
  • Select a metric: Upstream Latency (avg)
  • Group by: Service
  • Choose Granularity: Minutely

Then they add a filter to filter by the individual Runtime Group

  • Filter by: Runtime Group
  • Operator: In
  • Value: prod

Production - Upstream Latency (last hour)

Figure 2: Line chart showing the upstream latency over the last hour. The line chart shows the Account API spiking while the Payment API remains at normal levels.

Conclusion

With the two reports, Tal realizes that the Account API is having latency spikes that are impacting the overall response time. Because the platform has a latency of around 0.57ms, Tal can rule out the platform as the source of the problem. In the reports, Tal sees a small second spike in the last hour. In combination with the Latency API latency spikes, Tal will be able to use this to investigate the matter further.