Monitoring load balanced resources?
Posted: Thu Dec 29, 2011 6:32 pm
Hi there,
we have a number of webservers (www1-3) serving several websites (example1.com, example2.com, example3.com) via a loadbalancer (lb1). We also have a redundant loadbalancer (lb2). During normal operation, all websites are accessed via lb1, but in the event of a failure of lb1, they're all transparently migrated to serve via lb2. What is the best recommended way of validating operation-- the way I've configured it is:
1. www1, www2, www3 are each defined as hosts and their specific services are checked (ssh, mail, diskspace etc). They're all part of the 'webservers' hostgroup.
2. lb1 and lb2 are each defined as hosts with their specific services checked (ssh, mail, diskspace etc). They're both part of the 'loadbalancers' hostgroup.
3. I've created a 'pseudo host' called ha_websites with an address of 127.0.0.1 and then have a check_http (with the hostname of ha_websites) to verify that the websites are all accessible example1.com, example2.com etc.
To me this doesn't really feel ideal since glancing at the dashboard I can't tell which lb the websites are running on (beyond knowing that they'd be on lb2 if lb1 was down). Also, setting the ha_websites host address to 127.0.0.1 seems like a cheat. Whats the recommended way of doing this?
Thanks!
we have a number of webservers (www1-3) serving several websites (example1.com, example2.com, example3.com) via a loadbalancer (lb1). We also have a redundant loadbalancer (lb2). During normal operation, all websites are accessed via lb1, but in the event of a failure of lb1, they're all transparently migrated to serve via lb2. What is the best recommended way of validating operation-- the way I've configured it is:
1. www1, www2, www3 are each defined as hosts and their specific services are checked (ssh, mail, diskspace etc). They're all part of the 'webservers' hostgroup.
2. lb1 and lb2 are each defined as hosts with their specific services checked (ssh, mail, diskspace etc). They're both part of the 'loadbalancers' hostgroup.
3. I've created a 'pseudo host' called ha_websites with an address of 127.0.0.1 and then have a check_http (with the hostname of ha_websites) to verify that the websites are all accessible example1.com, example2.com etc.
To me this doesn't really feel ideal since glancing at the dashboard I can't tell which lb the websites are running on (beyond knowing that they'd be on lb2 if lb1 was down). Also, setting the ha_websites host address to 127.0.0.1 seems like a cheat. Whats the recommended way of doing this?
Thanks!