Am restartat serverul, apoi am rulat:     select * from sys.dm_os_performance_counters    si am obtinut:

Pentru primul counter  "Buffer cache hit ratio":

object_name                            Counter_name          Instance_name         cntr_value         cntr_type
SQLServer:Buffer Manager   Buffer cache hit ratio                                       2905              537003264

Apoi, in Sql Server Agent-Alerts  creez o alerta pe care am 'scriptat-o'(Script alert as-Create to) mai jos:

USE [msdb]
GO
/****** Object:  Alert [Buffer Manager: Buffer Cache Hit Ratio]    Script Date: 01/27/2010 10:42:27 ******/
EXEC msdb.dbo.sp_add_alert @name=N'Buffer Manager: Buffer Cache Hit Ratio',
        @message_id=0,
        @severity=0,
        @enabled=0,
        @delay_between_responses=0,
        @include_event_description_in=1,
        @category_name=N'[Uncategorized]',
        @performance_condition=N'SQLServer:Buffer Manager|Buffer cache hit ratio||<|90',
        @job_id=N'00000000-0000-0000-0000-000000000000'

Teoretic, acum alerta nu ar trebui sa semnalizeze nimic intrucat cntr_value=2905>90, insa ea returneaza mesajul(via email):

DESCRIPTION:    The SQL Server performance counter 'Buffer cache hit ratio' (instance 'N/A') of object 'SQLServer:Buffer Manager' is now below the threshold of 90.00 (the current value is 0.00).
Aici, chiar nu inteleg de unde ia aceasta valoare(=0.00) Pe alte servere unde am pus aceasta alerta sunt returnate valorile corecte din sys.dm_os_performance_counters(5000-6000)

Pentru al doilea counter "Average Latch Wait Time(ms)":

object_name                     Counter_name                          Instance_name         cntr_value         cntr_type
SQLServer:Latches    Average Latch Wait Time (ms)                                              351971          1073874176

Similar, creez in Sql Server Agent-Alerts, o alerta pe care o descriu mai jos:

USE [msdb]
GO
/****** Object:  Alert [Latches: Average Latch Wait Time]    Script Date: 01/27/2010 11:03:28 ******/
EXEC msdb.dbo.sp_add_alert @name=N'Latches: Average Latch Wait Time',
        @message_id=0,
        @severity=0,
        @enabled=1,
        @delay_between_responses=43200,
        @include_event_description_in=1,
        @category_name=N'[Uncategorized]',
        @performance_condition=N'SQLServer:Latches|Average Latch Wait Time (ms)||>|30000',
        @job_id=N'00000000-0000-0000-0000-000000000000'

Treshold-ul l-am fixat la 30secunde, iar alerta semnaleaza depasirea prin mesajul(via email):

DESCRIPTION:    The SQL Server performance counter 'Average Latch Wait Time (ms)' (instance 'N/A') of object 'SQLServer:Latches' is now above the threshold of 30000.00 (the current value is 351971.00).

Deci in acest caz, valoarea a fost corect luata din sys.dm_os_performance_counters