3 Replies Latest reply: Jun 26, 2014 1:56 PM by user3002968 RSS

    Alert Log Monitoring through OEM 12c

    Sreerag T R

      Dear All,

       

      We have enabled alert log monitoring through OEM 12c. To reduce the count of alerts, I would like to add exception in alerting.

       

      Please let me know how to stop alerting of non critical specific ORA- errors scuch as ORA-00060: Deadlock detected.

       

      Thanks in Advance..

       

      Regards

      Sreerag T R

        • 1. Re: Alert Log Monitoring through OEM 12c
          Loc Nhan -Oracle

          Hi Sreeraj,

           

          Assuming you are referring to metric Generic Alert Log Error, you can either change the regular expression in the warning/critical thresholds so it does not match the ORA- errors that you don't care about, or use the Alert Log Filter to filter out those errors. Please see the following MOS note for more information:

           

          Database Alert log monitoring in 12c explained (Doc ID 1538482.1)

           

          Regards,

          - Loc

          • 2. Re: Alert Log Monitoring through OEM 12c
            Sreerag T R

            Thanks Loc.

             

            From  the (Doc ID 1538482.1) it was mentioned to add the below expression in Alert Log Filtering field.

             

            .*ORA-0*(54|1142|1146|600:.*\[12240[^\]]*\])\D.*

             

            I'm eager to know how it works. Can anyone explain logic behind this.

             

            Thanks in Advance!!

             

            Sreerag T R

            • 3. Re: Alert Log Monitoring through OEM 12c
              user3002968

              Hello Sreerag,

               

              In OEM 12c, Oracle have introduced the DB Alert Log metric. If you looked in all metrics, going to Monitoring => Metric and Collection Settings, and change to All Metrics, you will see a metric called "Alert log" and another "DB Alert log".

               

              1-) "DB Alert Log" 

              "DB Alert Log" is a new one developed by Oracle, and it will only page you in case Oracle open a Incident report thru the ASR. For example, you have the following categories

               

              Archiver Hung Alert Log Error

              Reporting: ORA-00257 and ORA-16038

              Data Block Corruption Alert Log Error

              Reporting: ORA-01157, ORA-01578 and ORA-27048

              Session Terminated Alert Log Error

              Reporting: ORA 00603

              Media Failure Alert Log Error

              Reporting: ORA 01242 and ORA-01243

              Generic Alert Log Error

              All other alert log errors

               

              But this "DB Alert Log" is limited to a list of ORA-, because it only will create an Incident inside the OEM in case its really a critical issues, which create a Incident in ASR, that could be upload to Oracle support.

              The list of ORA- you can find executing the below query:

               

              SELECT * FROM V$DIAG_CRITICAL_ERROR;


              2-) "Alert log"

              This one is the old alert log of Oracle, basically follow the same behavior of DB Alert Log, however, now you don't have limitation of the ORA- to be reported. It contains the same list of categories, but the algorithm is not that good, you will receive a lot of Pagers and Alerts, one for each different error, etc.

               

              Conclusion:

               

              Basically what Oracle suggest and recommend you to do:

               

              1-) Use the "DB Alert Log";

              2-) Use the additional "Alert Log", in case your business/application require you to monitor a NON-critical issue (not included in V$DIAG_CRITICAL_ERROR)

              - Remember to enable the collector, will be disabled;

              - To avoid dupplication between "Alert Log" and "DB Alert Log", you have to click in the icon MULTIPLE PENCIL of Category "Generic Alert Log Error", from "Alert Log" metric and replace the content of "Filter Expression"


              ".*(ORA|OCI)-0*(54|1142|1146|600|603|7445|4030|4031|227|239|240|255|353|355|356|445|494|1578|3113|3137|4036|24982|25319|29740|29770|29771|32701|32703|32704|56729|3106|3135)*\D.*"


              Basically, right now, Oracle will use 2 ways of monitoring, one for real criticals and others that YOU and YOUR business think its critical.