We have been using Zenoss for the past year and we like it for the most part but there are some issues that have prompted us to look at other solutions. We have approximately 100 devices that are monitored today including: Windows servers, SQL, Apache, JBoss, Zimbra, HP Procurve, Astaro, Linux servers, etc. Zenoss does a great job at graphing network devices/hosts as well as Linux servers. It doesn't do as good with applications such as JBoss, Zimbra, Apache, etc. although there is a growing set of Zenpacks (like plugins) to extend monitoring, and the fact that we don't have to manage 100 agents is a plus to us since Zenoss uses SNMP and WMI.
I'm testing Hyperic HQ right now and we have a sample of our devices already loaded. So far the Windows and Linux server monitoring seems good but we're not so sure about what we are seeing with our Astaro firewall which is a Linux-based appliance, and an HP Procurve switch. We've added these as new platforms where the Astaro is a Network Host and the switch is a Network Device. We don't see any "Linux" performance data in HQ for the Astaro as we do in Zenoss and we don't see much either for the Procurve. The community string is correct for both devices and the other HQ information seems correct as well. In Zenoss we see load average, CPU, memory, disk I/O, all network interfaces, disk space, and more for our Astaro appliances and we see everything we care about for our Procurve switches as well.
We're concerned that Hyperic HQ is not going to allow us to monitor and graph our devices that don't support the agent. We have other devices that we need to monitor where we only care about whether the device is responding to pings or not, and we have still other hosts that we only want to monitor URLs on (i.e. check that the HTTP response is good and pass a username and password, then check for 'Welcome to ACME' on the returned page, etc.).
Can Hyperic HQ do all of this and are we missing something? I've posted a screenshot of the configuration for our network switch (an HP Procurve 5406). Again, in Zenoss, this device is discovered and graphed completely. In Hyperic HQ, we see no interfaces, no CPU, not much of anything really. Same is true for our Astaro firewall which Zenoss also discovers and graphs completely.
I'm testing Hyperic HQ right now and we have a sample of our devices already loaded. So far the Windows and Linux server monitoring seems good but we're not so sure about what we are seeing with our Astaro firewall which is a Linux-based appliance, and an HP Procurve switch. We've added these as new platforms where the Astaro is a Network Host and the switch is a Network Device. We don't see any "Linux" performance data in HQ for the Astaro as we do in Zenoss and we don't see much either for the Procurve. The community string is correct for both devices and the other HQ information seems correct as well. In Zenoss we see load average, CPU, memory, disk I/O, all network interfaces, disk space, and more for our Astaro appliances and we see everything we care about for our Procurve switches as well.
We're concerned that Hyperic HQ is not going to allow us to monitor and graph our devices that don't support the agent. We have other devices that we need to monitor where we only care about whether the device is responding to pings or not, and we have still other hosts that we only want to monitor URLs on (i.e. check that the HTTP response is good and pass a username and password, then check for 'Welcome to ACME' on the returned page, etc.).
Can Hyperic HQ do all of this and are we missing something? I've posted a screenshot of the configuration for our network switch (an HP Procurve 5406). Again, in Zenoss, this device is discovered and graphed completely. In Hyperic HQ, we see no interfaces, no CPU, not much of anything really. Same is true for our Astaro firewall which Zenoss also discovers and graphs completely.