Root-Cause analysis in case of API polling failure

There can be cases where the OpenStack services are running fine, but the polling of the OpenStack API may fail. To detect the unresponsiveness of the various API, the following set of events and problem are defined on Cloud Controller. These fault conditions are detected during polling only.

Fault condition, event name, and associated problems lists the set of events and problems that are defined on CloudController.

Table 1. Fault condition, event name, and associated problems

Fault condition	Event name	Problem
Response for the NOVA API is not received by ESM, however none of the application realizes that the NOVA service is down.	ComputeAPIServiceBroken	CloudControllerCommunicationBroken
Response for the Neutron API is not received by ESM, however none of the application realizes that the Neutron is down.	NetworkAPIServiceBroken	CloudControllerCommunicationBroken
Response for the CINDER API is not received by ESM, however none of the application realizes that the CINDER service is down.	BlockStorageAPIServiceBroken	CloudControllerCommunicationBroken