10 Replies Latest reply on Feb 20, 2019 7:09 AM by UmaSharma

    OS and Hardware Health Monitoring

    UmaSharma

      Hi Team,

      Can you please list down various mandatory counters to monitor health of OS and Hardware via Windows Performance Counters?

      It would be of great help!!

       

      Thanks

        • Re: OS and Hardware Health Monitoring
          Lal_Babu_Shaik

          Hi Uma

           

          Using below counters you can determine health status of your server.

           

          Disk Space : Select All the Drive letters which you want to monitor.

           

           

          Memory  :

           

          CPU : Select all Total or individual cores depending on your requirement

           

          System Uptime in System :

           

            • Re: OS and Hardware Health Monitoring
              UmaSharma

              Hi Lal Babu,

              Shall i consider only the highlighted once for monitoring Windows health?

              Could you elaborate a bit that why should we consider these highlighted counters? and rest of other counters will not serve the purpose?

              Thanks!

                • Re: OS and Hardware Health Monitoring
                  UmaSharma

                  Since i need to write analytics for them and try for possible combinations to determine each of its health. So I just wanted to be sure for example..'Available Mbytes' will help me or not in determining the there are any memory issues or not?

                    • Re: OS and Hardware Health Monitoring
                      Lal_Babu_Shaik

                      Hi Uma

                       

                      Disk Space : When you click on Free megabytes it will provide all drive on the server. You can create individual counter for each drive to monitor free space. e.g. If you are storing log files or IIS logs on any specific drive then it will grow day by day and space will be reduced. It is always good to monitor Archive drives ,backup drive freespace to avoid issues with low disk space.

                      Memory : Available memory provides utilisation. You can use this counter to check memory utlisation on the server. If load is high or computation is high then you can see using this counter.

                      CPU : You can monitor overall CPU against individual cores. Please note you can create individual counter for each core but overall will help you than creating individual. 

                      Both Memory and CPU can tell you if there is something wrong with utilisation on the server or server is busy processing requests.

                      System up time : It will be useful to check when server is rebooted. You can ignore if you want to use powershell or cmdlet to check the same.

                        • Re: OS and Hardware Health Monitoring
                          UmaSharma

                          Thanks ..this info is going to help a lot..Do you have any analytics written in AF for the above? I just wanted to have an idea since we want trigger event whenever there is an issue in OS

                            • Re: OS and Hardware Health Monitoring
                              Lal_Babu_Shaik

                              Hi Uma

                               

                              Request to follow below process.

                              1) Create PI Points for performance monitoring

                              2) Create a service account to run performance monitoring interface

                              3) Add service account to Performance monitoring group on each server.

                              4) Map this service account to Identity which has access to PI Points you want to write.

                              5) Create PI Perfmon Interface and map points to this interface. Define scan class accordingly with 1min, 5 min, etc based on your requirement.

                              6) Create AF hierarchy for each server. If you are having multiple server of same type e.g. PI collective then create a template with all attributes

                              7) Create elements from template and map PI Points if you are unable to use substitution in template

                              8) Use AF categories to separate Application parameters with Server paramaters in Template if required.

                              9) You can define limits on each attribute e.g. Lo,LoLo ,Hi,HiHi etc. You can may your original data to Minimum and Maximum e.g. if your server is allocated with 100 GB on C drive then map Minimum=0 ,Maximum=100 , Lo=30 and LoLo=20.

                              10) Create Enumeration set for status and define states accordingly .This will be helpful to map output from analysis.

                              11) If you are planning to create AF analysis for overall status then e.g. you have following attribute names

                              AF server Health : GOOD or BAD

                              AF Analysis : Running i.e. Analysis Running > 0

                              Buffer Subsystem: OK or Error etc.

                              Then sample equation : IF  (’AF server Health’ = “Good” and ’Analysis Running’>0 and ‘Buffer Subsystem’=“OK”) Then 1 Else 0
                              Map to output tag called Status mapped to enumeration set with 1=GOOD and 0= BAD (enumeration set)

                              12)use PI Vision or PB to create display for monitoring. You can use multi states defined in step 9 to show individual parameters and use output status tag for overall status.

                              13) Analysis defined in Step 11 can be used to set EF start trigger e.g. AF server health =“BAD” and end trigger  will be =“Good” based on your requirement you can add it.

                              14) You can dfine notification template and assign all attributes in the notification email body. You can also add link to PI Vision display or snapshot depending your requirement.

                               

                              Defined high level activies above And would request you to use based on your client requirement.

                              1 of 1 people found this helpful
                                • Re: OS and Hardware Health Monitoring
                                  UmaSharma

                                  Since for windows monitoring you have highlighted only one counter in each of the object like disk,memory etc.

                                  I think analytics for above will be straight forward and not a combination of 2 or 3 attributes( IF  (’AF server Health’ = “Good” and ’Analysis Running’>0 and ‘Buffer Subsystem’=“OK”) Then 1 Else 0 ). Pls correct me if i am wrong?

                                  How do you determine that we might need to have a combination of 2 or 3 attributes to calculate overall health of server?

                                  Do u think in my scenario which is just to know the server health i might want to write complex analysis to monitor the health?

                                    • Re: OS and Hardware Health Monitoring
                                      Lal_Babu_Shaik

                                      E.g. If you are running AF service, Analysis one server then above equation helps. You can define single analysis for each counter or combine depending on your requirement and use case.

                                       

                                      Another example : If Available Memory is less then LoLo then server is heavily loaded and you can configure accordingly to send email.

                                       

                                      On Server request you to open performance monitor and click each counter to check sub paramter e.g. disk free space will list all and you can select all .It is not a single parameter . It has sub parameters like C:,D:,E: Etc. If you archive is configure in D drive then it would be good to minotor D drive on the server and if space reduces to LoLo e.g. 10% of total available then notify so that you have time to get space allocated or do some clean up on the server or move archives to different drive.

                                      1 of 1 people found this helpful
                                        • Re: OS and Hardware Health Monitoring
                                          UmaSharma

                                          I am just wondering shall i include counters private bytes,Elapsed time,working set , %processor time of process ? Do they help in monitoring OS health? Since in other sites these are mentioned as important windows performance counters but i dn knw whther these will be helpful monitoring Windows OS or hardware health or not. 

                          • Re: OS and Hardware Health Monitoring
                            jyi

                            If you would like to monitor your IT Asset using PI System, perhaps below YouTube and toolkit might be helpful!

                            IT Asset Monitor Toolkit

                            Introduction to the IT Asset Monitor Toolkit

                            1 of 1 people found this helpful