Summary of Heartbeat HealthChecks and Application Monitoring mechanisms

book

Article ID: 167906

calendar_today

Updated On:

Products

XOS

Issue/Introduction

This article discusses and compares the purpose of the two health check mechanisms that verify the status of an Application Processor Module.This article discusses and compares the purpose of the two health check mechanisms that verify the status of an Application Processor Module.

Cause

This article discusses and compares the purpose of the two health check mechanisms that verify the status of an Application Processor Module.

Resolution

Heartbeat is a health check mechanism to determine the state of the system connections between modules on the control plane and data plane and verify whether these modules can communicate with one another. 

The CPM maintains the state of all connections in the system. If a module stops responding to heartbeats, the CPM automatically reloads the module in an attempt to recover from this condition. This mechanism is not user-configurable.
 
Application monitoring directs XOS to poll the application health on each VAP in a VAP group every five seconds to verify that the application is able to process traffic. It calls the file /crossbeam/apps/app_status on the VAP . This file is installed together with the application on the VAP group. By default, application monitoring is enabled and activated when an application is installed. If the application is down on a VAP, the XOS health monitoring system notifies the NPM to stop new flows to that VAP. The application monitor can be enabled or disabled for each VAP group from the XOS comand line. In addition, XOS can be instructed to reload a VAP whenever the application running on the VAP experiences a failure:
 
CBS# configure vap-group <VAP_group_name> [no] application-monitor [no] reload-on-app-failure [[no] wait-time <wait_time>]

Both of these health check mechanisms verify the status of the APMs, but at different levels (hardware and OS versus application processes). 
 

Workaround

N/A