Monitor and Analyze Nodes

The Nodes page displays all nodes linked to a selected cluster in the Manage Clusters page. This page provides real-time metrics to help you analyze node health and detect issues early.

Steps

1. In the Pulse UI, select Nodes from the left navigation bar > Summary.

2. Or, go to Pulse Home and click Overview.

3. On the Nodes page, set the time range and refresh the status:

   1. Set time range – Select Today, Last 12 hours, Last 3 months, or choose a custom period. Then select Apply.
   2. Refresh status – Select Play (⏵) to refresh the page automatically every 10 seconds. Select Pause (⏸) to stop auto-refresh.

4. On the Nodes page, you can analyze the following metrics for each node.

Node Details

• Host name: Click to view detailed performance data for the node or host. For details, Visualize Node Details.
• Host type: Identifies whether the node is a compute node or a data node.

Service Status

• Application / Service status: Number of services currently down; click the status icon to view connected (up) or disconnected (Down) services.
• Services Count: Running services compared to the total running on the node.

Resource Usage

• System load: Average number of tasks waiting in the kernel run queue; helps detect overload.
• CPU Usage (%): Percentage of CPU resources in use; useful for spotting high usage.
• Memory Usage (%): Percentage of memory resources in use; helps identify bottlenecks.
• Disk performance (read/write time): Total Time spent on disk read/write operations; indicates storage performance.

Workload Distribution

• YARN containers: Number of active containers running on the node; helps track workload distribution.

Analyze with the Heatmap

On the Nodes page, you can monitor node resource usage and performance using the heatmap.

Expand a Node

The heatmap provides a visual overview of a node’s performance metrics over time. Each colored block represents a metric’s value during a specific time interval.

• Color intensity: Darker shades indicate higher utilization or activity; lighter shades indicate lower values.
• Time intervals: Hover over a block to see the exact value and timestamp for that metric.

Metrics Tracked in the Heatmap

• CPU Usage (%): Shows the processing power used by the node. Darker blocks indicate higher CPU usage.
• Load: Represents the system load (the number of processes waiting for CPU time). Darker shades indicate higher load.
• Memory Usage (%): Shows the percentage of memory in use. Darker shades indicate higher memory utilization.
• Number of Containers: Displays the number of YARN containers running.

How to Use the Heatmap

• Spot Trends: Identify periods of high CPU, memory, or load.
• Correlate Metrics: Link spikes in containers with CPU or memory usage.
• Identify Anomalies: Detect sudden dark patches indicating unexpected load or resource pressure.