Health Monitor making more requests to target servers than expected

Hi

PFB configuration of the load balancer with Health monitor(HTTP) in place.

<IntervalInSec>60</IntervalInSec> 

is configured to be 60 seconds. but we could see almost 19k request being hit to our target servers

in an hour.

What would be cause of surge in these many requests???

<HTTPTargetConnection>
<LoadBalancer>
<Algorithm>RoundRobin</Algorithm>
<Server name="Price-promotions-1a"/>
<Server name="Price-promotions-1b"/>            
<Server name="Price-promotions-1c"/>            
<MaxFailures>5</MaxFailures>            
<RetryEnabled>true</RetryEnabled>        
</LoadBalancer>        
<Properties/>        
<Path>/promotions</Path>        
<HealthMonitor>            
<IsEnabled>true</IsEnabled>            
<IntervalInSec>60</IntervalInSec>            
<HTTPMonitor>                
<Request>                    
<ConnectTimeoutInSec>10</ConnectTimeoutInSec>                    
<SocketReadTimeoutInSec>30</SocketReadTimeoutInSec>                    
<Port>443</Port>                    
<Verb>GET</Verb>                    
<Path>/promotions/_status</Path>
<Header name="TraceId">price:promotion:HM</Header>
                
</Request>                
<SuccessResponse>                    
<ResponseCode>200</ResponseCode>                 
</SuccessResponse>            
</HTTPMonitor>        
</HealthMonitor>    
</HTTPTargetConnection>

Regards,

Venkat

0 4 329
4 REPLIES 4

What version of Apigee Edge are you using?

@Shailendra Baxi - do you have any insight here?

Hi Dino

We are on 19.01.14.

Regards,

Venkat

@Venkat Tummala,

The number of Health Monitor requests made to a specific target server depends on some of the following factors:

  1. Number of Target Endpoints/Service Callout policies in which you have configured the health monitor for the specific backend server host
  2. Number of API Proxies in which you have configured the health monitor for the specific backend server host
  3. Number of environments in which you have configured the health monitor for the specific backend server host
  4. Number of Message Processors you have configured across all regions/data centers you have for the specific org/environment
  5. How often (interval) are these Health Monitor requests being made ?

So you need to calculate the expected number of Health Monitor requests made in an hour based on the above factors. If the expected number does not match or is much lower/higher than the actual number of Health Monitor requests, then please reach out to Apigee Support.

Not applicable

There is no other parameter than the number of message pprocessors and number of APIs hitting the backend which can lead to this. So, please cross verify.

I would suggest to check the source of the request hitting the backend, whether those are from the message processor ips or not and also do count how many requests from each message processor.