When a microservice experiences latency spikes or elevated error rates, you need to determine which callers are responsible. The Upstream application tab on the Instance Details page tracks response time, request count, and error count for each application that sends data to your service. Use these metrics to isolate the upstream caller contributing to performance degradation.
This feature requires Professional Edition or Platinum Edition. Standard Edition does not support upstream application monitoring.
View upstream application metrics
To view upstream application metrics for a specific instance, complete the following steps:
Log on to the EDAS console.
In the left-side navigation pane, choose Application Management > Applications.
In the top navigation bar, select a region. In the upper part of the page, select a microservice namespace.
From the Cluster Type drop-down list, select Kubernetes Cluster.
Click the target application name in the Application Name column.
In the left-side navigation pane of the application details page, choose Monitor > Instance Details.
In the left-side pane of the Instance Details page, select an instance, then click the Upstream application tab.
The tab displays three time series charts for the selected time period:
| Chart | Description |
|---|---|
| Response Time | Response time of calls from each upstream application to your application. A sudden increase for a specific caller indicates that caller may be sending problematic requests. |
| Requests | Total request count from each upstream application. Use this chart to identify traffic distribution across callers and detect unexpected traffic spikes. |
| Errors | Number of failed requests from each upstream application. Compare error counts across callers to locate the source of failures. |
Interact with the charts
All three charts support the following interactions:
View details: Hover over a data point to see exact values at that timestamp.
Zoom in on a time range: Click and drag across a section of the chart to zoom into that period.
Diagnose upstream issues
After you identify an anomaly in the upstream application charts, follow these steps to investigate:
Correlate metrics: Check whether a response time spike for a specific upstream caller coincides with an increase in its request count or error count. A simultaneous spike in requests and response time suggests the caller is overloading your service.
Narrow the time range: Drag across the anomaly period in any chart to zoom in and inspect the exact timing.