I'm actually sort of flabbergasted that there is no way to do this.
Let's say I have 500 Linux systems currently being monitored by Hyperic. Then let's say we are told that we must start running a new daemon process on each of these systems and must monitor that it's always running.
There's no way for me to add, say, a process check to all these platforms automatically? I have to go to each individual platform and add the service? Even in EE?
Thanks for the pointer. I will say that this honestly is not that intuitive. You'll have to excuse my aggravation. I fought over a year and a half to get Hyperic into our infrastructure and you can ask Javier about how painful it was if you need any proof. Now that we're trying to implement it, we're just finding some shortcomings that are surprising.
However, the fact that it's taken this long to implement something as simple as global additions of a service check to a platform and in fact has to be done via scripting, to me, is... interesting?
One of the reasons we wanted to migrate from something like Nagios to Hyperic was that Nagios is a pain to administer, but even it has the ability to quickly add a check to all hosts of a group.
Historically, HQ has not been easy to make global modifications to. I understand your aggravation.
However, we've recently made a commitment to Groovy and the HQU plugin framework to alleviate any of these shortcomings. You now have full power over doing global modifications to your entire infrastructure. "It's just a simple matter of programming"
I know that's not a sexy answer, but people have very diverse needs and this is a great way that we can support all of them.
If you prefer the big sexy, HQU also allows you to easily create screens and UI elements. If you find that you have a frequent operation that you're performing, investigate creating some custom UI for it and submit it back.
We are excited about what kinds of things people are doing with the framework, so feel free to post any questions or things you've come up with.
Well, it's not that it's programming that bothers me - it's that it's java programming. 🙂
Java falls alot more into that 'oooo, scary, it's a 'real' programming language' sorta category. Where as scripting languages such as perl, python and such are more the sys admin's forte and requiring sys admins to learn, albeit not java, but a close derivative, just in order to obtain some global functionality in the 'Enterprise' version seems somewhat backwards.
I realize that by introducing HQU and the cool API, you provide people the ability to provide their own solutions to problems and the many customizations that the Hyperic dev team simply can't take on, it just seems there are some things that should have already been in the Enterprise version that make it much more, well, enterprise.
I appreciate your help and the pointers in the right direction Jon.