Resource Monitoring
Resource monitoring provides visual dashboards for hosts, middleware, and Kubernetes clusters. The left submenu is displayed dynamically according to the enabled types in Data Sources.
Submenus
| Submenu | Data Source |
|---|---|
| Host | node_exporter |
| MySQL | mysqld_exporter |
| MongoDB | mongodb_exporter |
| Redis | redis_exporter |
| Kafka | kafka_exporter |
| Elasticsearch | elasticsearch_exporter |
| Kubernetes | kube-state-metrics/kubelet. See Kubernetes Cluster Monitoring |
| Custom pages | Any external Web UI (such as Flink), registered in Data Sources |
Host Monitoring
The "Host" page summarizes resource status for all hosts. The top section is a resource detail table for all hosts (CPU/memory/load/disk/partition usage, zombie processes, health score, and more). The lower section expands time-series details for CPU, memory, disk, and network by host. Click a host IP to drill down into the full host dashboard.

Middleware Monitoring
Each middleware type has a dedicated dashboard. Use the "Instance" dropdown at the top to switch among multiple instances of the same type.
MongoDB
Runtime duration, QPS, latency, replica set role, command operations, connections, query efficiency, and more. The top contains "Monitoring" and "Slow Query" tabs. See MongoDB Slow Query Analysis for the latter.

MySQL
Uptime, QPS, connections, InnoDB buffer pool, client thread activity, command statistics, and more.

Redis
Uptime, connections, memory usage, command execution/hit rate, network I/O, key counts by DB, and more.

Kafka
Message production/consumption rate, consumer group lag, topic partition distribution, and more.

Elasticsearch
Cluster health, shard status, node CPU/JVM heap memory, disk and network, document indexing rate, and more.

Dashboard Operations
Resource monitoring dashboards are embedded from Grafana. Common operations:
- Instance selector: Switch among multiple instances of the same type from the
Instancedropdown in the upper-left corner. - Time range: Switch the time window in the upper-right corner (default: last 3 hours/30 minutes). Custom ranges and auto-refresh are supported.
- Expand/collapse: Click group row titles in the dashboard to expand or collapse them.
Text in Grafana custom dashboards is consistently in English and does not change with the platform language. English dashboard titles and legends do not affect metric data.
Kubernetes and Custom Pages
Kubernetes: Monitors cluster nodes, containers, Pods, and cluster object status (node resources, namespaces, CPU/memory quotas, Pod distribution, and more). For collection methods, see Kubernetes Cluster Monitoring.

Custom pages: External Web UIs registered on the "Data Sources" page (such as Flink or Langfuse) appear in the "Custom" sidebar group. In reverse proxy mode, the Ops Platform forwards requests; in direct embed mode, the browser connects to the target address. See Data Sources for configuration.