Kevin Holman's System Center Blog

Posts in this blog are provided "AS IS" with no warranties, and confers no rights. Use of included script samples are subject to the terms specified in the Terms of UseAre you interested in having a dedicated engineer that will be your Mic

Monitoring ConfigMgr 2012 with OpsMgr

Monitoring ConfigMgr 2012 with OpsMgr

  • Comments 22
  • Likes

 

As with almost all Microsoft products, there is an MP to monitor ConfigMgr 2012.  This MP is such a divergence from the previous ConfigMgr 2007 MP, I wanted to take a moment and point out the differences.

 

  • Not a conversion MP!

The previous ConfigMgr 2007 MP was actually a conversion MP from MOM 2005.  This caused major issues in supporting that management pack, and there were a long list of things that needed “fixing” along the way, which resulted in required updates to the MP.  The ConfigMgr 2012 MP was written from the ground up alongside ConfigMgr 2012.  It plugs directly into the registry and WMI to connect to the built in health analyzer in ConfigMgr.  Because it is not a conversion MP, there is no longer a dependency on the backwards compatibility MP. 

  • NO SCRIPTS in the Monitoring MP! 

The previous MP has about 12 scripts, which fed hundreds of rules which generate alerts.  These scripts were problematic, and left very few options for override tuning.  Now the MP primarily uses shared datasources which leverage WMI or registry data.

  • Lines of code

The precious MP had 125,000+ lines of code in the XML.  The new MP has less than 40,000.  That is a huge reduction.

  • Number of workflows

The previous MP had over 900 rules.  The current MP has 214 rules, and only ONE is enabled out of the box!  All the monitoring comes from Monitors now.

  • Disabled workflows out of the box.

Out of the box, the 2012 MP ships with ALL performance collection rules disabled.  This allows the customer to turn these on if they desire all this performance data in their console or for reporting.  This lessens the impact of the MP on the monitored servers, and the SCOM infrastructure.  Additionally, there are many monitors that are disabled out of the box, due to uncommon monitoring scenarios, un-deployed roles, or potential noise.  The guide neatly calls these out.

  • A well documented guide and appendix.

The current MP ships with a REALLY well documented MP guide that explains what is necessary to get the MP working, and an appendix which outlines every class, and detailed information about each discovery, rule, or monitor, including data source info.

  • Simpler health rollup and class

This thing is so simple – its beautiful!

 

 

image

 

 

 

Here is a breakdown on the MP:

 

 

What does it do???  (Key Monitoring Scenarios)

 

  • Replication Health
    • SQL replication between sites
    • Primary > CAS > Primary
    • Primary to Secondary
    • Receiving and Sending
  • Replication Configuration
    • Service broker validity
    • SQL port
    • SQL firewall port
    • SQL disk space
    • SQL DB certificate validity
  • Backup and Recovery status
  • Component Availability
    • Inbox Manager
    • Site component manager
    • Management Point
    • Outbox
    • WSUS
    • Site System status
    • Deployment Status
    • Component status
  • Service Availability
    • Alert on critical service status
      • SMSexec
      • SMS Agent Host
    • Dependent services
      • WMI
      • SQL
      • IIS
      • BITS
      • WSUS
  • Server Role Availability
    • Asset intelligence
    • Management Point
    • Reporting Point
    • Software Updates
    • Distribution point
    • Application catalog
    • Enrollment point
    • Fallback status
    • Endpoint protection
  • Backlog monitoring (DDR, Inventory)
  • General Health (Process/Processor)
  • Database connection status
  • Software update synchronization (WSUS)
  • DP Configuration

 

 

 

What’s in the MP?

Microsoft.SystemCenter2012.ConfigurationManager.Library.mp

  • 13 Discoveries
  • 84 Classes
  • 14 Groups

Microsoft.SystemCenter2012.ConfigurationManager.Discovery.mp

  • 13 Discoveries
  • 2 Classes

Microsoft.SystemCenter2012.ConfigurationManager.Monitoring.mp

  • 145 Monitors
  • 214 Rules (all but one disabled)
  • 199 Monitoring Views
  • 34 Reports

 

 

 

What’s required from a configuration perspective?

 

Not much! 

  • Make sure your agents run as Local System
  • Enable agent proxy for site servers
  • Create an override MP for your tuning specific to this MP.

Optional configuration:

  • Check out the disabled rules and monitors documented in the guide and see if you want them enabled.
  • 52 performance threshold monitors that might need to be adjusted for your environment (see guide)
  • 15 manual reset monitors.  Optionally disable these and enable 15 rules which are corresponding to these, if your organization cannot support manual reset monitors (most cannot)

 

Distribution

Summary:

 

Looks like a solid MP, very simple to deploy, configure, and LIGHT YEARS better than that ConfigMgr 2007 MP.  Very nice!

 

With the new dashboarding capabilities in OpsMgr – creating a single pane of glass for the ConfigMgr team was never simpler:

image

Comments
  • Hmm..All my Secondary CM-sites in View "Sites" in OM2012 Console in state "Not monitored". Primary site - "Healthy"

  • ...And empty Views "Distribution Points - State","Management Points - State","Software Update Points - State"

  • And another strange thing. I get RED status from "SQL Server Firewall Port Monitor", but on all my CM-servers Port 1433 and 4022 is open :(

  • Which "•15 manual reset Monitors" do you mean? In the documentation all Monitors have Auto Reset behavior.

  • We have the same issue as Aleksey - Only the central site shows up as healthy, everything else as not monitored. SCOM 2012 SP1 monitoring SCCM 2012 SP1.

    Could this be caused by our agents running on a service account (with admin privileges on the server but not in sccm)?

  • Here is one more to Alexey and Jon.

    I do not agree that the MP is well-documented and simple.

    ConfigMgr 2007 on OpsMgr 2007 was painless, really.

  • I'm really missing documentation on the permission requirements for running the agents on a service account. I actually thought running the SCOM-agents on a domain service account was best practice and as many environments are set up that way it's not something you want to stray away from on a few servers if you don't must.

    I'd really like to se documentation on what privileges the action account needs. If possible with least privileges necessary within SCCM to avoid separation of duties-conflicts.

  • When we just imported the 2012 Management Pack we cant get this monitor back to green state:

    Site Database Server Without Valid Certificate (Database Certificate Validity Monitor)

    But we have a valid Certificate, so i dont know what this monitor reacts on.

    This info is gathered from Alert Context on the Alert:

    < DataItem type =" MOM.RegistryData " time =" 2013-05-29T15:15:55.0774480+02:00 " sourceHealthServiceId =" 79B707F7-327F-214F-02FD-8523D5BB9141 " >

    < Values >

     < Severity VariantType =" 19 " > 3 </ Severity >

     < State VariantType =" 19 " > 3332 </ State >

     </ Values >

     </ DataItem >

    Now what?

  • Same as the guys above.  Fresh install of SCCM 2012 MP and no distribution points have been discovered.  The all server state view shows these as "Not Monitored"

    all agents running as localsystem with agent proxy enabled.

  • No update on a possible solution for this, I'm in the same boat; primary SCOM agents healthy, SCCM MP Server and Site System Roles showing as 'Not Monitored' 24 hours after enabling them as agent proxy. No active or closed alerts related to the servers either.

  • UPDATE!! Logged a support call for this after spending a fair bit of time troubleshooting, MS was stumped too and went off to confer with the MP owner, got back to me next dayt stating that the MP guide is incorrect and that you should enable proxy agent on the CAS and Primary site! I can confirm that discovery worked after that and all objects and health states are reporting correctly. Nice diagram too :-) Hope this helps!

  • Why I see rollup monitors in critical state, but all child unit monitors in healthy state? Reset monitor and maintenance mode not helps me.

  • @bobgreen84 -

    That can happen when there is a crash on the management server responsible for calculating health state.  I generally try maintenance mode on the affect entity, or flush the health service cache on all the management servers.

  • Kevin, out of curiosity, do you have an 'automated process' to flush the health service cache or are you referring to the manual rename method? Thanks.

  • Same issue.  Not only does it not discover DPs, but my site server is showing under "Servers and Site System Roles" as "Not Monitored".  I have renamed the "Health Service State" folder on the client to clear the cache, but not luck yet.  I'm about to remove the MP and re-add.  I couldn't find anything listing Prerequisite MPs needed, but it would make sense to require SQL and IIS.  Last time I used this MP, I added it after SQL and IIS.  It worked with no issue.  It should really state in the guide if there are prerequisite MPs, but I can't find it.  I didn't read the whole doc, but i'm a pretty good skimmer.

    -LvilleSystemsJockey

Your comment has been posted.   Close
Thank you, your comment requires moderation so it may take a while to appear.   Close
Leave a Comment
Search Blogs