Background
Network faults are among the most common issues in Kubernetes, and troubleshooting them can be challenging, often requiring intervention from specialized Ops personnel. To address this, the CFG provides five types of network fault actions: network disorder, network duplication, network package loss, network latency, and network DNS anomalies. These actions allow you to independently stabilize network performance in your business and ensure the normal operation of services.
Experiment Implementation
Experiment Preparation
Purchase container instances in a standard cluster and deploy test services.
Experiment Steps
Step 1: Create an Experiment
2. Click Skip and create a blank experiment, and fill in the experiment details.
Step 2: Add Experiment Instances and Actions
Node Type Network Faults
Note:
Applicable resource objects: Regular node in standard cluster.
1. During the experiment object configuration, select the container object type Standard Cluster Node.
2. Add a Regular Node in Standard Cluster node instance. Select the Cluster ID where the fault will be injected, and the node under that cluster will be automatically added.
3. Add experiment actions. Select Network latency fault action under network resources.
4. Configure the action parameters.
Pod Type Network Faults
Note:
Applicable resource objects: Standard cluster Pod and Serverless cluster Pod.
1. Select Standard Cluster Pod as the container object type.
2. Add a Pod instance, then sequentially select the required Cluster ID> Load Type (Workload) > Namespace> Load Name (Workload) to select a Pod instance that meets the criteria.
3. Select Network latency fault action under network resources.
4. Configure fault action parameters. On this page, you can configure whether to enable Container faults. If it is enabled, after you enter the container name, the fault can be injected into the specified container.
Step 3: Configure Monitoring (Optional)
In the Global Configuration stage, you can select monitoring metrics, check the necessary ones, and add them to the configuration.
Step 4: Execute the Experiment
Click Execute Experiment. Since network actions are high-risk, authentication must be completed before executing the action.