Device Summary
Introduction
Device Summary is a function designed for array servers to view key indicators of array sub-nodes. The page queries historical monitoring data by device and time range, and displays key indicators such as CPU, memory, temperature, network, disk, fan, NPU and power supply in curve charts.
Development Vision
- Provide administrators with a unified multi-node performance trend view to quickly identify resource usage anomalies and mutations.
- Supports query by time range, sampling step and refresh cycle, while meeting the needs of real-time inspection and historical trend analysis.
Implementation logic
aBMC queries Prometheus historical data based on the selected node, start and end time, and Step, and combines each node and device indicator into a paging and hideable trend curve.
Feature Usage
View and analyze equipment operating trends
Set query conditions
- Select Devices in the main navigation bar on the left.
- Select Summary in the secondary navigation bar.
- Select the aBMC or subnode to be queried in the device selector. The page supports selecting multiple nodes at the same time, and all enabled nodes are selected by default.
- Select the query time range, for example Last 30 Min.
- Set Step within the recommended range displayed on the page.
- Select the automatic refresh interval, optional
15 Seconds,30 Seconds,1 Minute,3 Minutesor5 Minutes. - When you need to get the latest data immediately, click the refresh button.
- Reset Data is used to reset historical monitoring data, not a chart refresh button.
Step Settings
Step represents the sampling interval of trend data, in seconds. After the time range is changed, the page will recalculate the recommended values and available upper and lower limits. The smaller the value, the denser the data points; the larger the value, the sparser the data points.
Clear monitoring data carefully
Reset Data will reset the Prometheus monitoring database and affect the historical trends of all nodes, instead of just clearing the current chart or the currently selected device. Data retention requirements should be confirmed before execution and only performed by authorized administrators.
View basic resources and download traffic trends
After setting the query conditions, the page automatically refreshes all trend charts. Start by looking at the chart at the top of the page:
- CPU Usage: View the CPU usage trend of each node.
- Memory Usage: View the memory usage trend of each node.
- Temperature: View the CPU or SoC temperature trend of each node.
- Network download speed: Check the downlink speed of each node and each network card.
When comparing multiple nodes, first confirm the node name through the legend, and then observe the difference in the curves at the same time point.
View upload, processor and disk trends
Scroll down the page and view the images below:
- Check the network card uplink speed in Network upload speed.
- Check the running frequency of each CPU core in CPU Frequency.
- Check the read rate of each storage device in Disk Read Data Rate.
- Check the write rate of each storage device in Disk Write Data Rate.
View fan, NPU and power supply power trends
Continue to scroll to the bottom of the page and check the hardware health status as shown below:
- View the real-time speed of each fan in Fan Speed.
- Check the NPU usage of each node in NPU Usage.
- View device power supply trends in Power Supply Wattage.
No monitoring data
When the corresponding hardware is not installed on the current node, no indicators are reported, or there are no collection records within the selected time range, the chart will display No Data Available.
Analyze chart data
- Move the mouse into the chart to view the indicator values of the corresponding time points, nodes and devices.
- Click the legend item to show or hide the curve of the corresponding node, network card, CPU core or storage device.
- When there are many legend items, use the arrows on both sides of the legend to switch paging.
- After changing the device, time range or step, wait for all charts to finish refreshing, and then check the timeline and legend.
Trend chart indicator description
| Chart | Description | Common units |
|---|---|---|
| CPU Usage | CPU usage. | % |
| Memory Usage | Memory usage. | % |
| Temperature | CPU or SoC temperature. | ℃ |
| Network download speed | Network card download speed. | The page automatically converts bps, Kbps or Mbps according to the value |
| Network upload speed | Network card uplink speed. | The page automatically converts bps, Kbps or Mbps according to the value |
| CPU Frequency | The operating frequency of each core of the CPU. | GHz |
| Disk Read Data Rate | Storage device read rate. | The page automatically converts B/s, KB/s or MB/s according to the value |
| Disk Write Data Rate | Storage device write rate. | The page automatically converts B/s, KB/s or MB/s according to the value |
| Fan Speed | Fan real-time speed. | RPM |
| NPU Usage | NPU usage. | % |
| Power Supply Wattage | Equipment power supply. | W |
FAQ
Chart display No Data Available
First confirm that you have selected the correct node and time range, then click the refresh button. If only a certain type of chart has no data, it may be that the node does not have the corresponding hardware, or the current version does not report the corresponding Prometheus indicators.
The chart data points are too dense or too sparse
Adjust Step within the recommended range of the page. Smaller steps can be used for short time ranges, and larger steps are recommended for long time ranges.
Some nodes or equipment curves are not displayed
Check the top device selector and legend status. If the legend is hidden, click the corresponding legend item to redisplay the curve; when there are many legends, use the arrows on both sides to switch to the corresponding page.
The chart is not updated for a long time
Click Manual Refresh and check the automatic refresh interval, node online status, and management network. When a node is offline or Prometheus collection is abnormal, the historical curve may be interrupted or stopped updating.
5.Fan Speed continues to display No Data Available
Confirm whether the device supports fan speed collection and check with the fan presence status and RPM on the System > Thermal page.
The historical trend disappears after Reset Data
Reset Data will reset the Prometheus historical database, and the cleared data cannot be restored through pages. New data needs to wait for each node to continue collecting before it reappears.