In case of Synchronous API requests, you can make Adeptia Connect to first check if the runtime pod has either enough CPU, or memory, or both available to be considered ready to accept any more request. The application performs this check by using the following two configurable threshold properties for runtime.
- readiness.probe.cpu.threshold
- readiness.probe.memory.threshold
| Property | Description |
|---|---|
| readiness.probe.cpu.threshold | Threshold value (in percentage of the allocated CPU request) for CPU utilization. If the pod CPU utilization goes beyond this value, the pod will not accept any more requests to process. Ensure that this threshold value is always greater than the HPA target CPU utilization percentage value that has been set through the environment variable RUNTIME_AUTOSCALING_TARGETCPUUTILIZATIONPERCENTAGE in the global values.yaml file. You can keep the property value blank to skip this threshold check. |
| readiness.probe.memory.threshold | Threshold value (in percentage of the minimum JVM memory XMS) for application memory utilization. If the application memory utilization goes beyond this value, the pod will not accept any more requests to process. Ensure that this threshold value is always greater than the HPA target memory utilization percentage value that has been set through the environment variable RUNTIME_AUTOSCALING_TARGETMEMORYUTILIZATIONPERCENTAGE in the global values.yaml file. You can keep the property value blank to skip this threshold check. |
You can find these properties at Account >Settings >Microservice Settings >Runtime >Readiness Probe Configuration. To know how to configure these properties, refer to the page Configuring the application properties.
Related Topics