-
Multi-tenant?
I’m not 100% sure that multi-tenant is the right terminology (although I think it is), but basically I wanted to see if there was a way to use Pandora FMS to monitor multiple different clients at different locations. I’m an IT service provider, and wanted to offer monitoring to my clients, but obviously the data and monitoring has to be kept entirely separate since I’m dealing with multiple different clients. Is that something that Pandora FMS can do? If so, any pointers/guidance on how to do it?
Thanks!
-
10 Replies
-
Hello
It is possible. You need to establish the Pandora FMS server at a public ip address. Install agents in the clients machines and make them be directed to the server. I believe you will need the Enterprise versión of Pandora FMS for your needs. If you’d like more information on the Enterprise version you can contact us at: info@artica.es
Regards
Ivo -
-
mbressman, why do you need to silo the data? It’s not like you’ll be collecting L.O.B. data — are you? I too am a service provider and see no issues concerning monitor collected data. I’m at week two, in my Pandora FMS community edition lab trial, and besides a few small kinks – and the learning curve – I think it’ll do just fine for replacing my current commercial remote monitoring solution. I’ve used groups to provide visual separation between clients and their sites and, though it may be too soon to judge, I think that is the way to accomplish separation.
-
mbressman, why do you need to silo the data? It’s not like you’ll be collecting L.O.B. data — are you? I too am a service provider and see no issues concerning monitor collected data. I’m at week two, in my Pandora FMS community edition lab trial, and besides a few small kinks – and the learning curve – I think it’ll do just fine for replacing my current commercial remote monitoring solution. I’ve used groups to provide visual separation between clients and their sites and, though it may be too soon to judge, I think that is the way to accomplish separation.
Good point – I guess you’re saying that since I’m the only one with access (and potentially anyone else working for me that I give access to), there’s no need to keep the data separate, right?
Would love to get some insight into how you are set up, and what small kinks you’ve encountered? And also what commercial remote monitoring solution you previously used (although feel free to PM me if you don’t want to post it on here).
Also, don’t mean to sound like an idiot here, but what is L.O.B. data?
Thanks for the post and the info!
-
L.O.B. == Line Of Business
Kinks:
1. getreg not working, had to write batch files containing reg query commands
2. logevent modules not working – still pursuing this rabbit (probably going to use wevutil.exe)
3. an agent running on SBS 2011 wasn’t using UTF-8 encoding, though it was configured to, which caused the Pandora server to not parse the returned XML file (now fixed)
4. Checking services from an agent running on SBS 2011 causes wmiprvse.exe to crash (think this is the cause/resolution but you’ve got to register for the hotfix (http://support.microsoft.com/kb/981314) and I haven’t been able to because the website backend system is malfuntioning)Regarding configuration, I am using the appliance in the lab, running on Hyper-V 2008 R2 (full Windows Server) but plan to roll out within a Server 2012 Hyper-V production environment if the lab trial goes well. I need to monitor between 200 and 300 nodes, ranging from Dell Servers, Desktops, laptops, and PowerConnect switches, Dell SonicWALL TZ and NSA appliances, HP Procurve switches, and various brands of printers and MFDs.
The goal is to monitor using software agents for every thing except gateway router/firewalls, where
– servers report back every 15 minutes
– desktops/laptops report in every hour
– Core switches and gateways report back every 5 minutes for availability checks but everything else every 15-30 minutes
– printers and MFDs report in every hour, mainly just error and usage info, as supplies are handled by the customersSince I believe you’re kind of new to all of this, I’ll tell you, make sure your customers use products designed to be monitored and managed. Consumer and SOHO wares will absolutely ruin your chances of providing a managed service of any nature.
-
L.O.B. == Line Of Business
Kinks:
1. getreg not working, had to write batch files containing reg query commands
2. logevent modules not working – still pursuing this rabbit (probably going to use wevutil.exe)
3. an agent running on SBS 2011 wasn’t using UTF-8 encoding, though it was configured to, which caused the Pandora server to not parse the returned XML file (now fixed)
4. Checking services from an agent running on SBS 2011 causes wmiprvse.exe to crash (think this is the cause/resolution but you’ve got to register for the hotfix (http://support.microsoft.com/kb/981314) and I haven’t been able to because the website backend system is malfuntioning)Regarding configuration, I am using the appliance in the lab, running on Hyper-V 2008 R2 (full Windows Server) but plan to roll out within a Server 2012 Hyper-V production environment if the lab trial goes well. I need to monitor between 200 and 300 nodes, ranging from Dell Servers, Desktops, laptops, and PowerConnect switches, Dell SonicWALL TZ and NSA appliances, HP Procurve switches, and various brands of printers and MFDs.
The goal is to monitor using software agents for every thing except gateway router/firewalls, where
– servers report back every 15 minutes
– desktops/laptops report in every hour
– Core switches and gateways report back every 5 minutes for availability checks but everything else every 15-30 minutes
– printers and MFDs report in every hour, mainly just error and usage info, as supplies are handled by the customersSince I believe you’re kind of new to all of this, I’ll tell you, make sure your customers use products designed to be monitored and managed. Consumer and SOHO wares will absolutely ruin your chances of providing a managed service of any nature.
Thanks for the reply and information – it is greatly appreciated. I’m beginning to think that Pandora FMS might not be the right product for me. Basically, I have a bunch of small businesses and individual/home users that I want to offer monitoring services to, but pretty much all of them (with a few exceptions) use consumer and SOHO hardware and products. Unfortunately, there’s no way any of them would upgrade just to get monitoring (or for any reason pretty much), so I’m stuck with trying to fit my square peg (monitoring services) into their round holes (consumer/SOHO equipment) – hopefully that doesn’t come off as an awful pun!
Any ideas if Pandora FMS will work in this scenario, and if not maybe you could point me to another product or recommend one that might be better suited?
Thanks!
-
I’ve used a fair number of monitoring solutions, commercial and FOSS, and Pandora FMS ranks up there with the best of them. Whether it is right for you is entirely based on your preference but yes, the community edition will work … though it sounds like you haven’t made much of an effort to try for yourself. You’ll need to setup a pandora server and learn how to configure the local software agents. You can also perform simple monitoring of the edge routers, as long as they allow ping.
I’ve move my pandora project out of the lab phases and am now piloting, using a test group of 80 servers, routers, and workstations and, besides the things I’ve already mentioned, it has been going rather well. The thing is, it’s going to cost you something, either time or cash (maybe both).
-
I’ve used a fair number of monitoring solutions, commercial and FOSS, and Pandora FMS ranks up there with the best of them. Whether it is right for you is entirely based on your preference but yes, the community edition will work … though it sounds like you haven’t made much of an effort to try for yourself. You’ll need to setup a pandora server and learn how to configure the local software agents. You can also perform simple monitoring of the edge routers, as long as they allow ping.
I’ve move my pandora project out of the lab phases and am now piloting, using a test group of 80 servers, routers, and workstations and, besides the things I’ve already mentioned, it has been going rather well. The thing is, it’s going to cost you something, either time or cash (maybe both).
True – I haven’t made much effort but figured it made sense to first post on here and find out more information before spending time only to later find out it wouldn’t do anything I wanted. Now that I have a better idea though, I plan to give it a shot on a VirtualBox implementation – does that make sense? Also, do you think it will work OK with consumer/SOHO equipment? And do you happen to know it can monitor other things, like websites for instance?
As far as it costing me something (time or money) – what did you mean by this? I figured I’d have to spend time learning how it works initially, but just planned to throw it on an Ubuntu server I have and didn’t think I’d necessarily have to spend any additional money. Am I wrong?
Thanks.
-
You should really spend a few days reviewing the documentation. Pandora can monitor almost anything, given you have access to whatever and whatever is equipped to be monitored with standard techniques.
Here’s my advice:
1. List your requirements
2. List your expected issues
3. Review the quick start guides, but read the admin guide
4. Run the appliance, as a VM, during lab trials (I spent about 3-4 weeks in the lab)
5. Search the docs and then use this forum, for specific instructions/questions regarding your current tasks and those immediately to follow
6. Once satisfied with the lab results, move to a new installation for piloting a pool of test subjects (I am currently in week 3, I think, would really have to check back into the project notes)
7. After successfully piloting, with 90% of desired functionality achieved, move to a production environment. If you’re careful not to make a mess during piloting, the pilot system can be cloned for production use; that way you still have a test environment after you go live.
