Skip to main content

Execution & Monitoring

Combine all stress factors into a single 2-minute benchmark and observe the results via the centralized monitoring server.

Benchmark Monitoring Topology​

1. Combined Benchmark Command​

To perform a holistic stress test on the client nodes, execute the following combined command:

sudo stress-ng --cpu 0 --cpu-load 60 --vm 1 --vm-bytes 30% --hdd 1 --hdd-bytes 1G --timeout 120s

stress-ng combined

2. Server-Side Monitoring​

While the benchmark is running on the clients (10.12.12.22 and 10.12.12.23), navigate to the Server Monitoring (Grafana) dashboard to verify the utilization spikes:

  1. Open Grafana Dashboard.
  2. Select the Node Exporter Full dashboard.
  3. Observe the CPU, Memory, and Disk I/O graphs during the 120-second window.

grafana compare

warning

Ensure that the Node Exporter and Promtail agents are active on the clients to capture the performance data during the test.