How to reduce health monitor interval when pool member is down?

Hey guys,

 

We've got an HTTP health monitor currently checking a pool of web service hosts. It seems that when a pool member is restarted, we see 'bad packet header' errors in either app logs, likely due to the incoming health checks before the web service is fully online/healthy. So we're thinking it may be good to tune the health monitor interval if possible.

 

Currently, we're using a 5s/16s int/timeout for this request: GET / HTTP/1.0\r\n\r\nConnection: Close\r\n

 

We see very occasional TCP reset messages (maybe three each day total, from either LTM in the HA pair). But the 'bad packet header' seems to happen during app restarts where the service is not fully up yet.

 

I see the 'up interval' advanced health monitor setting, which allows me to specify a different check interval while the web service is up. The app admin prefers we use a 30s interval while the pool member is down, and we maintain a 5s interval while the pool member is up. So I was thinking I could change the monitor as follows:

 

Interval: 30s Up Interval: 5s Timeout: 16s

 

Does anyone have any thoughts or experience doing this type of health monitor tweaking?

 

Thanks in advance! B