We have NPM,SAM and NMP Integrated VMM in our environment. We need to monitor as much as possible. I find though that this product suite requires a huge amount of manual work and I just don't have time. What I'd lie to do is the following: * discover all VMs running in ESXi and add them as nodes to NPM with polling for VMWare and ICMP status only. Since we have VMM integrated we don't need NPM do do anything but we do need SAM to do things and same needs the target in as a node in NPM. How to do this automatically? * For all new physical servers that come on the network how to discover and put an agent on automatically and poll using agent? * How to automatically detect IIS,Sharepoint,Exchange etc. on detected hosts

Community
- Command Central
- MVP Program
- Monthly Mission
- Blogs
- Groups
- Events
- Media Vault
Products
- Observability
- Network Management
- Application Management
- IT Security
- IT Service Management
- System Management
- Database Management
Content Exchange
- SolarWinds Platform
- Server & Application Monitor
- Database Performance Analyzer
- Server Configuration Monitor
- Network Performance Monitor
- Network Configuration Manager
- SQL Sentry
- Web Help Desk
Free Tools & Trials
Store

Automate Discovery and Monitoring

We have NPM,SAM and NMP Integrated VMM in our environment. We need to monitor as much as possible. I find though that this product suite requires a huge amount of manual work and I just don't have time. What I'd lie to do is the following:

discover all VMs running in ESXi and add them as nodes to NPM with polling for VMWare and ICMP status only. Since we have VMM integrated we don't need NPM do do anything but we do need SAM to do things and same needs the target in as a node in NPM. How to do this automatically?
For all new physical servers that come on the network how to discover and put an agent on automatically and poll using agent?
How to automatically detect IIS,Sharepoint,Exchange etc. on detected hosts

Find more posts tagged with

Accepted answers

All comments

jamesatloop1

You cant deploy an agent automatically through solarwinds. Your only option is setting up an automatic scan and import while only adding wmi credentials and root credentials for the hosts. You will need to select poll for vmware also. To circumnavigate the agent issue, i would advise you to add the agent to your golden master image, that way all newly deployed vm's will automatically add to solarwinds. An auto discovery and and import will then add the appinsight templates to the new vm nodes. As for the physical servers, your stuck with manual deployment.

shocko

Thank for the reply! SO say I get my agent onto my systems (with ESD or whatever) how do you discover applications on them automatically? You mentioned:

An auto discovery and and import will then add the appinsight templates to the new vm nodes

Can you elaborate please?

orioncrack

You're still struggling with the lack of automation eh? I feel your pain.

Automatic Discovery of Resources to Monitor

shocko

Yes. I'm kind of fed up with this product in a lot of ways but I'm stuck with it. I'm using the SDK quite a bit but that's for another rant on another thread!

HerrDoktor

Why would you want to monitor as much as possible? Is anyone looking at those statistics/values/alerts? In most organizations the answer is no. So again, why monitor as much as possible? It puts a lot of stress on your monitoring machines and on your nerves. I would concentrate on the important stuff to monitor so you have enough time for:

- proper baselining

- proper alerting

- proper custom-monitoring

- building your own views

I see this in a lot of projects where my clients or prospects want to "see everything" so I can relate to your pain. But instead of solving how to force feed as much data into the system as possible. why not look at it from a different angle.

Make the application-/node-/device-owners tell you what they need to have monitored, and make it clear that "everything" is not a valid answer. I created templates with some of my clients to give to the stakeholders and fill out what they need, this works in some organizations, not in all.

shocko

Ok, then might you be able to outline a process for the following:

2000 servers, mix of Linux,Windows,Unix
60% production
Many using dynamically allocated storage using things like NetApp flexclones
Many DevOps systems that have services added on the fly (installs etc.)

Now how to get these systems into NPM/SAM automatically and get cores things like running services monitored automatically? I'm a very busy sysadmin and I don't have time to do this manually via a slow web interface.

In most orgs I would argue that (for compliance and other reason) they want to see everything then build production views. I appreciate taking the time to respond but I feel you need to flesh out your respond a little more to show how this can work in a mid to large size organization. Auto discovery is the core of a lot of other monitoring products.

HerrDoktor

Do you have some sort of CMDB where your devices/servers/... are listed?

some of my clients use FNT-Command for this. In these cases command is the "leading" System telling Solarwinds what to monitor with which credentials with which baselines. -> Unfortunately this is SDK Programming. My sync-Script has about 1500 Lines of code but does its job on a daily basis for about 4 years now. But from your last post I see that you would like to rant about the SDK so I assume this might not be an Option for you.

Others use a Sharepoint List where I can pull the data to put into solarwinds. -> Same strategy -> Import Scripts with API Integration

As a "busy system admin" you can't know everything so someone has to tell you what is important on which machines.

When talking about "core things" in your case I would go and use SNMP on the machines. It makes life easier for you, at least from my experience with dynamically allocated storage (all other connection options won't work so well when the disk-IDs change)

Let a "Network discovery" run with just a few options. Status / CPU / Memory so you have the Nodes in your system.

For your services: there is a "all automatic Services monitor" for Windows systems: Check All Services Which Set To Automatic Mode Are Running I am not sure if this works with SNMP but you might be able to specify WMI for this. Apply sometihing like this monitor as "default" to your windows machines so you have all that are set to "Automatic"

The AppInsight templates can be assigned by automatic discovery, however I have seen too many unnecessary IIS-Monitors going red so I would not automatically assign that template.

I am not very familiar with Linux or other UX systems so I would need more thinking and trying on that.

Point out to your Management that monitoring a midsize environment is not a task you do when you have other things to do. As an example: My projects with cliens start at 20 days for a very basic Solarwinds installation and consulting. The larger ones order about 40 consulting days per year for improvements and updates.

I have seen that you had SCOM /SCCM before, so what was the main reason for switching the monitoring product?

Cheers,

Holger

shocko

Thanks for the reply. I also, just to point out that I think myself and others are just trying to drive some change within the product rather than having a rant

SOCM lost traction for the following:

It was seen as hard to implement bu some admins
Didn't do network monitoring well
IN 2007 didn't really have strong dash boarding

mesverrum

It's fair to say that improving automatic discovery should be a big priority within solarwinds, because these days it is one of the biggest pain points I see with my clients. Almost all the required pieces exist and you can script up a pretty robust automation if you are strong in the sdk, so that tells me there is no reason the team at solarwinds couldn't just build more intelligence into it out of the box. For example a decently workable scheme I've used does auto discovery based on AD OU's or against defined subnets, then I set up scripts to parse out what it finds and remove/ignore the things I don't need, add in everything that is left, shortly after that it assigns templates based on whatever criteria we deemed were appropriate, typically I like to just pick apart the installed software list of the asset inventory and auto assign templates from there.

On the one hand, it is a strength of the solarwinds platform that you can crack it open and make it do almost anything you want, but on the other hand, that should not give them am excuse not to make the off the shelf product more useful on its own.

jamesatloop1

To get the templates to auto assign all you have to do is setup a group in manage groups and add new group. Add a dynamic query, now depending on your server naming\ip policies, say we will go with hostname, set the query to node, then select system name from the dropdown. Type in a naming convention such as 'sql' or 'web' if that is contained within the naming convention of your servers hostname and save, Your groups should be by server function such as iis, sql etc.... Then go to sam settings, manage templates and tick the specific template you want to auto assign. From the the tool bar select 'assign to group'. Going forward anything that gets added via the autodiscovery gets automatically filtered into your group nad has the template automatically assigned whether be it agent, wmi or ssh.

orioncrack

Its not this easy, nor reliable for very large environments.

Its a very kludgey process.

jamesatloop1

It is very annoying to set this up alright BUT you only have to do it once. I work with a client at the moment who has retail outlets all over the world which equates to 6 polling engines, 20,000 + devices and a TB database and it works very well for them.

jlhartsock

Second this comment. We've tried using groups for auto assignments of SAM monitors and have found very mixed results. Some groups work completely flawlessly and then others will have random members join and drop from the groups for no reason. I've went through multiple support cases and more diag captures than I could even imagine. At this point, it's just not reliable enough for our organization.