Microservices Engine:Monitoring center

Enable the Grafana monitoring dashboard

New Grafana monitoring dashboard

If you have enabled the Grafana monitoring dashboard, the new version offers a wider range of monitoring metrics.

1. Log in to the MSE console and select a region in the top navigation bar.

2. In the left-side navigation pane, choose Microservices Registry > Instances.

3. On the Instances page, click the name of your target instance.

4. In the left-side navigation pane, choose Monitoring Center.

   Note

   • By default, the dashboard displays data from the last 15 minutes. You can select a different time range or specify a custom one from the drop-down list in the upper-right corner.

   • Hover over a data point on a chart to view the data for a specific node at that time, with minute-level granularity.

   • Click the Refresh icon in the upper-right corner to update the monitoring data.

   • On the Overview tab, you can view the following information:

     • In the Overview section, view metrics such as Number of nodes, Number of configurations, Number of service providers, Queries per Second, Operations per Second, and Number of Connections.

     • In the Usage Watermark section, view configuration usage, service provider usage, and connection usage.

     Note

     If you use a Eureka client, connection-related metrics on the Overview tab cannot be collected because Eureka clients use short-lived connections.

   • Click the Registry Monitoring tab to view metrics such as Number of services, Number of service providers, Number of service subscribers, Registry TPS, Registry QPS, Registry Write RT, and Registry Read RT. TPS stands for transactions per second, QPS stands for queries per second, and RT stands for response time.

     Note

     • Nacos 2.0.4 and later versions provide four internal services for address discovery over the Application Configuration Management (ACM) protocol. Therefore, the values for Number of services and Number of service providers on the monitoring charts are four higher than the actual count. This is expected behavior.

     • If you use a Eureka client, the number of service subscribers cannot be collected because Eureka clients use polling queries instead of subscriptions.

   • Click the Configuration Center Monitoring tab to view metrics such as Number of configurations, Number of configuration listeners, Configuration Center TPS, Configuration Center QPS, Configuration Center Write RT, and Configuration Center Read RT.

     Note

     In the Monitoring Center dashboard, Configuration Center Monitoring represents configuration query QPS (queries per second) — the number of configuration read requests processed per second — rather than an independent transaction-processing metric. You can combine this metric with HTTP request counts to assess the overall load on the configuration center. Both real-time and historical monitoring data are available on this tab.

   • Click the Push Monitoring tab to view metrics such as Service Push Success Rate, service push latency, Service Push TPS, and empty push ratio.

     Note

     Eureka clients use polling queries, which do not trigger Nacos push functionality. As a result, push-related metrics cannot be collected.

   • Click the Connection Monitoring tab to view metrics such as Number of Client Versions and Number of long-lived connections.

     Note

     Because Eureka clients use short-lived connections, metrics on the Connection Monitoring tab cannot be collected.

   • Click the JVM Monitoring tab to view Java Virtual Machine (JVM) metrics such as Young GC Time, Young GC Count, Full GC Time, Full GC Count, and Heap Memory Usage. GC stands for garbage collection.

   • Click the Resource Monitoring tab to view resource metrics such as inbound traffic, outbound traffic, Memory Usage, CPU Usage, Number of nodes, and system load.

   • Click the TopN Monitoring tab to view the following information:

     • In the Service TopN Dashboard section, you can view Top N service providers, Top N service subscribers, and Top N IP Push Failures.

     • In the Configuration TopN Dashboard section, you can view Top N configuration changes and Top N configuration listeners.

5. (Optional) For more advanced observability features, click Import to Grafana Expert Edition in the upper-right corner. This redirects you to the Grafana console, where you can explore powerful, multi-tenant dashboard capabilities.

6. (Optional) To embed an engine monitoring page, click Open <Monitoring Name> in New Window in the upper-right corner to get its direct URL. <Monitoring Name> represents the name of the current monitor.

   For example, on the Registry Monitoring tab, click Open Registry Monitoring in New Window in the upper-right corner to open the registry monitoring page in a new browser tab.

Old dashboard

If you have not enabled the Grafana monitoring dashboard, you can continue to use the old dashboard. However, we recommend upgrading to the new Grafana dashboard to access a richer set of metrics. For instructions, see Enable the Grafana monitoring dashboard.

1. Log in to the MSE console and select a region in the top navigation bar.

2. In the left-side navigation pane, choose Microservices Registry > Instances.

3. On the Instances page, click the name of your target instance.

4. In the left-side navigation pane, click Observation Analysis.

5. On the Observation Analysis page, click the Monitor tab to view metrics such as Number of services, Number of service providers, and Average Response Time of Service Write API (ms).

   Note

   • By default, the dashboard displays data from the last 30 minutes. You can select other time ranges, such as Last 30 minutes, Last 1 hour, Last 6 hours, and Last 24 hours, or specify a custom time range.

   • The monitoring data for the three nodes in the engine is displayed in different colors. You can click a node name below a chart to show or hide its data. At least one node must remain visible in the chart.

   • Hover over a data point on a chart to view the data for all three nodes at that specific time, with minute-level granularity.

   • Click the Refresh icon in the upper-right corner to update the monitoring data.

FAQ

Q: The Redis RT displayed in Monitoring Center is abnormally high (for example, 3.759 seconds). Is this value accurate?

In most cases, an unusually high Redis RT value in the monitoring dashboard is a display artifact caused by the sampling method rather than the true Redis response time. To verify the actual Redis response time, check the request logs for your instance. Use the response time recorded in the request logs as the authoritative reference for Redis RT.

Q: Node restart or failover events appear in the monitoring data, but my business traffic is not affected. What is the cause?

When underlying anomalies occur in an MSE instance, the system may trigger single-node failover, restart, or migration operations. These events are reflected as anomaly indicators or alerts in the monitoring data. Because MSE Nacos engines use multi-node cluster deployment by default (3 nodes by default), a single-node failover is part of the normal high-availability mechanism and typically does not affect overall business availability.