@@CPU_BUSY

I'm looking at the various counter variables available in SQL Server 2000 &
2005 and trying to understand what they can be used for. The ones I'm
intrested in are:
@.@.cpu_busy
@.@.io_busy
@.@.idle
@.@.pack_received
@.@.pack_sent
@.@.connections
@.@.packet_errors
@.@.total_read
@.@.total_write
@.@.total_errors
Now I've read BOL on these and understand they get reset when restarting the
instance on SQL Server and I've also looked at sp_monitor and the
spt_monitor system table. However, I'm still fishing.
Were these intended to tell a DBA how busy the server has been?
Can they be used to determin when you have more capicity that you need?
How about that you don't have enough capicity to do the job you are trying
to?
Can I use them to determin jumps in activity after software has been
updated?
We have about 20 database servers in the farm and I would like to add some
kind of monitor to tell me which servers can be placed on a virtual server
without having to visit each one regurally as well as to have a historical
record.
Thanks,
JayHi Jay
In general perfmon is probably a better tool to do what you require as the
instantanious values given by the listed functions don't really have a
context. You may find the values more useful for before/after sampling to ge
t
an indication of what has been consumed, although there are also alternate
ways to do this, such as SQL Profiler.
John
"Jay" wrote:

> I'm looking at the various counter variables available in SQL Server 2000
&
> 2005 and trying to understand what they can be used for. The ones I'm
> intrested in are:
> @.@.cpu_busy
> @.@.io_busy
> @.@.idle
> @.@.pack_received
> @.@.pack_sent
> @.@.connections
> @.@.packet_errors
> @.@.total_read
> @.@.total_write
> @.@.total_errors
> Now I've read BOL on these and understand they get reset when restarting t
he
> instance on SQL Server and I've also looked at sp_monitor and the
> spt_monitor system table. However, I'm still fishing.
> Were these intended to tell a DBA how busy the server has been?
> Can they be used to determin when you have more capicity that you need?
> How about that you don't have enough capicity to do the job you are trying
> to?
> Can I use them to determin jumps in activity after software has been
> updated?
> We have about 20 database servers in the farm and I would like to add some
> kind of monitor to tell me which servers can be placed on a virtual server
> without having to visit each one regurally as well as to have a historical
> record.
> Thanks,
> Jay
>
>|||The problem with both Perfmon and SQL Profiler is that they require
connecting to the server. With 20 servers and growing, I would prefer to use
a data feed and write a report that summarizes the information and tells me
which servers need some attention. Since I already have a summary system
setup, I was hoping to add server usage stats to it, but only if they
provide me with some value.
Also, if Perfmod and the Profiler are better, why do these values exist?
"John Bell" <jbellnewsposts@.hotmail.com> wrote in message
news:451AC64E-6A91-46F4-AF8F-FFCCD1A03F0C@.microsoft.com...[vbcol=seagreen]
> Hi Jay
> In general perfmon is probably a better tool to do what you require as the
> instantanious values given by the listed functions don't really have a
> context. You may find the values more useful for before/after sampling to
> get
> an indication of what has been consumed, although there are also alternate
> ways to do this, such as SQL Profiler.
> John
>
> "Jay" wrote:
>|||Jay,
I don't have any suggestion on these functions (as these are called nowadays
) per se, since I don't
use them. I'd just like to give my view on your below thought:

> Also, if Perfmod and the Profiler are better, why do these values exist?
These are mostly a left-over since the very early days of the product. BEfor
e Windows NT (nowadays
"Windows"), SQL Server ran on OS/2. OS/2 didn't have Performance Monitor. So
these were what we had.
Note that we are talking about the years before 1992 here, so pretty ancient
(in the software
world). Of course, one could argue that MS should have deprecated these a lo
ng time ago, and for
that I don't have a good reply... :-)
Tibor Karaszi, SQL Server MVP
http://www.karaszi.com/sqlserver/default.asp
http://sqlblog.com/blogs/tibor_karaszi
"Jay" <nospam@.nospam.org> wrote in message news:uOklLmJKIHA.5224@.TK2MSFTNGP02.phx.gbl...[vbc
ol=seagreen]
> The problem with both Perfmon and SQL Profiler is that they require connec
ting to the server. With
> 20 servers and growing, I would prefer to use a data feed and write a repo
rt that summarizes the
> information and tells me which servers need some attention. Since I alread
y have a summary system
> setup, I was hoping to add server usage stats to it, but only if they prov
ide me with some value.
> Also, if Perfmod and the Profiler are better, why do these values exist?
>
> "John Bell" <jbellnewsposts@.hotmail.com> wrote in message
> news:451AC64E-6A91-46F4-AF8F-FFCCD1A03F0C@.microsoft.com...
>[/vbcol]|||Hi Jay
Perfmon is probably the tool you wish to use the most, this can be run (and
it is preferrable) on a remote server. The can be scheduled to run at given
periods or continuously run at a pre-defined sampling solution. Alerts can
also be scheduled if the perfmon value exceeds a defined limit. Other system
management tools such as MOM, Tivoli or HP Openview etc can also be used to
monitor your systems and they also provide additional functionality for a
more comprehensive approach to system monitoring.
John
"Jay" wrote:

> The problem with both Perfmon and SQL Profiler is that they require
> connecting to the server. With 20 servers and growing, I would prefer to u
se
> a data feed and write a report that summarizes the information and tells m
e
> which servers need some attention. Since I already have a summary system
> setup, I was hoping to add server usage stats to it, but only if they
> provide me with some value.
> Also, if Perfmod and the Profiler are better, why do these values exist?
>
> "John Bell" <jbellnewsposts@.hotmail.com> wrote in message
> news:451AC64E-6A91-46F4-AF8F-FFCCD1A03F0C@.microsoft.com...
>
>

@@CPU_BUSY

I'm looking at the various counter variables available in SQL Server 2000 &
2005 and trying to understand what they can be used for. The ones I'm
intrested in are:
@.@.cpu_busy
@.@.io_busy
@.@.idle
@.@.pack_received
@.@.pack_sent
@.@.connections
@.@.packet_errors
@.@.total_read
@.@.total_write
@.@.total_errors
Now I've read BOL on these and understand they get reset when restarting the
instance on SQL Server and I've also looked at sp_monitor and the
spt_monitor system table. However, I'm still fishing.
Were these intended to tell a DBA how busy the server has been?
Can they be used to determin when you have more capicity that you need?
How about that you don't have enough capicity to do the job you are trying
to?
Can I use them to determin jumps in activity after software has been
updated?
We have about 20 database servers in the farm and I would like to add some
kind of monitor to tell me which servers can be placed on a virtual server
without having to visit each one regurally as well as to have a historical
record.
Thanks,
Jay
Hi Jay
In general perfmon is probably a better tool to do what you require as the
instantanious values given by the listed functions don't really have a
context. You may find the values more useful for before/after sampling to get
an indication of what has been consumed, although there are also alternate
ways to do this, such as SQL Profiler.
John
"Jay" wrote:

> I'm looking at the various counter variables available in SQL Server 2000 &
> 2005 and trying to understand what they can be used for. The ones I'm
> intrested in are:
> @.@.cpu_busy
> @.@.io_busy
> @.@.idle
> @.@.pack_received
> @.@.pack_sent
> @.@.connections
> @.@.packet_errors
> @.@.total_read
> @.@.total_write
> @.@.total_errors
> Now I've read BOL on these and understand they get reset when restarting the
> instance on SQL Server and I've also looked at sp_monitor and the
> spt_monitor system table. However, I'm still fishing.
> Were these intended to tell a DBA how busy the server has been?
> Can they be used to determin when you have more capicity that you need?
> How about that you don't have enough capicity to do the job you are trying
> to?
> Can I use them to determin jumps in activity after software has been
> updated?
> We have about 20 database servers in the farm and I would like to add some
> kind of monitor to tell me which servers can be placed on a virtual server
> without having to visit each one regurally as well as to have a historical
> record.
> Thanks,
> Jay
>
>
|||The problem with both Perfmon and SQL Profiler is that they require
connecting to the server. With 20 servers and growing, I would prefer to use
a data feed and write a report that summarizes the information and tells me
which servers need some attention. Since I already have a summary system
setup, I was hoping to add server usage stats to it, but only if they
provide me with some value.
Also, if Perfmod and the Profiler are better, why do these values exist?
"John Bell" <jbellnewsposts@.hotmail.com> wrote in message
news:451AC64E-6A91-46F4-AF8F-FFCCD1A03F0C@.microsoft.com...[vbcol=seagreen]
> Hi Jay
> In general perfmon is probably a better tool to do what you require as the
> instantanious values given by the listed functions don't really have a
> context. You may find the values more useful for before/after sampling to
> get
> an indication of what has been consumed, although there are also alternate
> ways to do this, such as SQL Profiler.
> John
>
> "Jay" wrote:
|||Hi Jay
Perfmon is probably the tool you wish to use the most, this can be run (and
it is preferrable) on a remote server. The can be scheduled to run at given
periods or continuously run at a pre-defined sampling solution. Alerts can
also be scheduled if the perfmon value exceeds a defined limit. Other system
management tools such as MOM, Tivoli or HP Openview etc can also be used to
monitor your systems and they also provide additional functionality for a
more comprehensive approach to system monitoring.
John
"Jay" wrote:

> The problem with both Perfmon and SQL Profiler is that they require
> connecting to the server. With 20 servers and growing, I would prefer to use
> a data feed and write a report that summarizes the information and tells me
> which servers need some attention. Since I already have a summary system
> setup, I was hoping to add server usage stats to it, but only if they
> provide me with some value.
> Also, if Perfmod and the Profiler are better, why do these values exist?
>
> "John Bell" <jbellnewsposts@.hotmail.com> wrote in message
> news:451AC64E-6A91-46F4-AF8F-FFCCD1A03F0C@.microsoft.com...
>
>

@@CPU_BUSY

I'm looking at the various counter variables available in SQL Server 2000 &
2005 and trying to understand what they can be used for. The ones I'm
intrested in are:
@.@.cpu_busy
@.@.io_busy
@.@.idle
@.@.pack_received
@.@.pack_sent
@.@.connections
@.@.packet_errors
@.@.total_read
@.@.total_write
@.@.total_errors
Now I've read BOL on these and understand they get reset when restarting the
instance on SQL Server and I've also looked at sp_monitor and the
spt_monitor system table. However, I'm still fishing.
Were these intended to tell a DBA how busy the server has been?
Can they be used to determin when you have more capicity that you need?
How about that you don't have enough capicity to do the job you are trying
to?
Can I use them to determin jumps in activity after software has been
updated?
We have about 20 database servers in the farm and I would like to add some
kind of monitor to tell me which servers can be placed on a virtual server
without having to visit each one regurally as well as to have a historical
record.
Thanks,
JayHi Jay
In general perfmon is probably a better tool to do what you require as the
instantanious values given by the listed functions don't really have a
context. You may find the values more useful for before/after sampling to get
an indication of what has been consumed, although there are also alternate
ways to do this, such as SQL Profiler.
John
"Jay" wrote:
> I'm looking at the various counter variables available in SQL Server 2000 &
> 2005 and trying to understand what they can be used for. The ones I'm
> intrested in are:
> @.@.cpu_busy
> @.@.io_busy
> @.@.idle
> @.@.pack_received
> @.@.pack_sent
> @.@.connections
> @.@.packet_errors
> @.@.total_read
> @.@.total_write
> @.@.total_errors
> Now I've read BOL on these and understand they get reset when restarting the
> instance on SQL Server and I've also looked at sp_monitor and the
> spt_monitor system table. However, I'm still fishing.
> Were these intended to tell a DBA how busy the server has been?
> Can they be used to determin when you have more capicity that you need?
> How about that you don't have enough capicity to do the job you are trying
> to?
> Can I use them to determin jumps in activity after software has been
> updated?
> We have about 20 database servers in the farm and I would like to add some
> kind of monitor to tell me which servers can be placed on a virtual server
> without having to visit each one regurally as well as to have a historical
> record.
> Thanks,
> Jay
>
>|||The problem with both Perfmon and SQL Profiler is that they require
connecting to the server. With 20 servers and growing, I would prefer to use
a data feed and write a report that summarizes the information and tells me
which servers need some attention. Since I already have a summary system
setup, I was hoping to add server usage stats to it, but only if they
provide me with some value.
Also, if Perfmod and the Profiler are better, why do these values exist?
"John Bell" <jbellnewsposts@.hotmail.com> wrote in message
news:451AC64E-6A91-46F4-AF8F-FFCCD1A03F0C@.microsoft.com...
> Hi Jay
> In general perfmon is probably a better tool to do what you require as the
> instantanious values given by the listed functions don't really have a
> context. You may find the values more useful for before/after sampling to
> get
> an indication of what has been consumed, although there are also alternate
> ways to do this, such as SQL Profiler.
> John
>
> "Jay" wrote:
>> I'm looking at the various counter variables available in SQL Server 2000
>> &
>> 2005 and trying to understand what they can be used for. The ones I'm
>> intrested in are:
>> @.@.cpu_busy
>> @.@.io_busy
>> @.@.idle
>> @.@.pack_received
>> @.@.pack_sent
>> @.@.connections
>> @.@.packet_errors
>> @.@.total_read
>> @.@.total_write
>> @.@.total_errors
>> Now I've read BOL on these and understand they get reset when restarting
>> the
>> instance on SQL Server and I've also looked at sp_monitor and the
>> spt_monitor system table. However, I'm still fishing.
>> Were these intended to tell a DBA how busy the server has been?
>> Can they be used to determin when you have more capicity that you need?
>> How about that you don't have enough capicity to do the job you are
>> trying
>> to?
>> Can I use them to determin jumps in activity after software has been
>> updated?
>> We have about 20 database servers in the farm and I would like to add
>> some
>> kind of monitor to tell me which servers can be placed on a virtual
>> server
>> without having to visit each one regurally as well as to have a
>> historical
>> record.
>> Thanks,
>> Jay
>>|||Jay,
I don't have any suggestion on these functions (as these are called nowadays) per se, since I don't
use them. I'd just like to give my view on your below thought:
> Also, if Perfmod and the Profiler are better, why do these values exist?
These are mostly a left-over since the very early days of the product. BEfore Windows NT (nowadays
"Windows"), SQL Server ran on OS/2. OS/2 didn't have Performance Monitor. So these were what we had.
Note that we are talking about the years before 1992 here, so pretty ancient (in the software
world). Of course, one could argue that MS should have deprecated these a long time ago, and for
that I don't have a good reply... :-)
--
Tibor Karaszi, SQL Server MVP
http://www.karaszi.com/sqlserver/default.asp
http://sqlblog.com/blogs/tibor_karaszi
"Jay" <nospam@.nospam.org> wrote in message news:uOklLmJKIHA.5224@.TK2MSFTNGP02.phx.gbl...
> The problem with both Perfmon and SQL Profiler is that they require connecting to the server. With
> 20 servers and growing, I would prefer to use a data feed and write a report that summarizes the
> information and tells me which servers need some attention. Since I already have a summary system
> setup, I was hoping to add server usage stats to it, but only if they provide me with some value.
> Also, if Perfmod and the Profiler are better, why do these values exist?
>
> "John Bell" <jbellnewsposts@.hotmail.com> wrote in message
> news:451AC64E-6A91-46F4-AF8F-FFCCD1A03F0C@.microsoft.com...
>> Hi Jay
>> In general perfmon is probably a better tool to do what you require as the
>> instantanious values given by the listed functions don't really have a
>> context. You may find the values more useful for before/after sampling to get
>> an indication of what has been consumed, although there are also alternate
>> ways to do this, such as SQL Profiler.
>> John
>>
>> "Jay" wrote:
>> I'm looking at the various counter variables available in SQL Server 2000 &
>> 2005 and trying to understand what they can be used for. The ones I'm
>> intrested in are:
>> @.@.cpu_busy
>> @.@.io_busy
>> @.@.idle
>> @.@.pack_received
>> @.@.pack_sent
>> @.@.connections
>> @.@.packet_errors
>> @.@.total_read
>> @.@.total_write
>> @.@.total_errors
>> Now I've read BOL on these and understand they get reset when restarting the
>> instance on SQL Server and I've also looked at sp_monitor and the
>> spt_monitor system table. However, I'm still fishing.
>> Were these intended to tell a DBA how busy the server has been?
>> Can they be used to determin when you have more capicity that you need?
>> How about that you don't have enough capicity to do the job you are trying
>> to?
>> Can I use them to determin jumps in activity after software has been
>> updated?
>> We have about 20 database servers in the farm and I would like to add some
>> kind of monitor to tell me which servers can be placed on a virtual server
>> without having to visit each one regurally as well as to have a historical
>> record.
>> Thanks,
>> Jay
>>
>|||Hi Jay
Perfmon is probably the tool you wish to use the most, this can be run (and
it is preferrable) on a remote server. The can be scheduled to run at given
periods or continuously run at a pre-defined sampling solution. Alerts can
also be scheduled if the perfmon value exceeds a defined limit. Other system
management tools such as MOM, Tivoli or HP Openview etc can also be used to
monitor your systems and they also provide additional functionality for a
more comprehensive approach to system monitoring.
John
"Jay" wrote:
> The problem with both Perfmon and SQL Profiler is that they require
> connecting to the server. With 20 servers and growing, I would prefer to use
> a data feed and write a report that summarizes the information and tells me
> which servers need some attention. Since I already have a summary system
> setup, I was hoping to add server usage stats to it, but only if they
> provide me with some value.
> Also, if Perfmod and the Profiler are better, why do these values exist?
>
> "John Bell" <jbellnewsposts@.hotmail.com> wrote in message
> news:451AC64E-6A91-46F4-AF8F-FFCCD1A03F0C@.microsoft.com...
> > Hi Jay
> >
> > In general perfmon is probably a better tool to do what you require as the
> > instantanious values given by the listed functions don't really have a
> > context. You may find the values more useful for before/after sampling to
> > get
> > an indication of what has been consumed, although there are also alternate
> > ways to do this, such as SQL Profiler.
> >
> > John
> >
> >
> >
> > "Jay" wrote:
> >
> >> I'm looking at the various counter variables available in SQL Server 2000
> >> &
> >> 2005 and trying to understand what they can be used for. The ones I'm
> >> intrested in are:
> >>
> >> @.@.cpu_busy
> >> @.@.io_busy
> >> @.@.idle
> >> @.@.pack_received
> >> @.@.pack_sent
> >> @.@.connections
> >> @.@.packet_errors
> >> @.@.total_read
> >> @.@.total_write
> >> @.@.total_errors
> >>
> >> Now I've read BOL on these and understand they get reset when restarting
> >> the
> >> instance on SQL Server and I've also looked at sp_monitor and the
> >> spt_monitor system table. However, I'm still fishing.
> >>
> >> Were these intended to tell a DBA how busy the server has been?
> >> Can they be used to determin when you have more capicity that you need?
> >> How about that you don't have enough capicity to do the job you are
> >> trying
> >> to?
> >> Can I use them to determin jumps in activity after software has been
> >> updated?
> >>
> >> We have about 20 database servers in the farm and I would like to add
> >> some
> >> kind of monitor to tell me which servers can be placed on a virtual
> >> server
> >> without having to visit each one regurally as well as to have a
> >> historical
> >> record.
> >>
> >> Thanks,
> >> Jay
> >>
> >>
> >>
>
>

*** Interesting chart problem ****

Hi:
Lets say for eg: you have LCD projectors that you can borrow at various
times in a day. I need to create a chart that shows the 24 hours (like 12AM,
1AM, 2Am, 3Am...11PM, 12PM), on the X axis, and the LCD resource name on the
Y axis, and if it was booked between 2AM and 4AM, have a colored bar between
2Am and 4Am, and rest blank, and so on for all the resources.
What are my options to do something like this. Cant seem to find any way to
do this.
Thanks
NIThere is no good way to achieve this kind of visualization - but here is an
idea: use a scatter line chart and create the dataset so that you have the
start time and the end time for one projector in different rows. Anyway, it
won't be easy to achieve this.
-- Robert
This posting is provided "AS IS" with no warranties, and confers no rights.
"Nagini Indugula" <NaginiIndugula@.discussions.microsoft.com> wrote in
message news:DFF5D09D-28B7-4DAB-B86C-5A2EDE25D978@.microsoft.com...
> Hi:
> Lets say for eg: you have LCD projectors that you can borrow at various
> times in a day. I need to create a chart that shows the 24 hours (like
> 12AM,
> 1AM, 2Am, 3Am...11PM, 12PM), on the X axis, and the LCD resource name on
> the
> Y axis, and if it was booked between 2AM and 4AM, have a colored bar
> between
> 2Am and 4Am, and rest blank, and so on for all the resources.
> What are my options to do something like this. Cant seem to find any way
> to
> do this.
> Thanks
> NI|||Thanks Robert for the solution. I got through the problem, by creating a
table, and then having 24 columns ( for each hour) in the table, and did some
math, and changed the background color of the cells in the table. With some
math I figured out whihc is the first cell in the table that needs to be
colored, and the last cell, and accordingly acheived a gantt chart look.
However it would be nice if RS includes a gantt chart option in the chart
options. (kind of like the MS project charts).
Thanks
"Robert Bruckner [MSFT]" wrote:
> There is no good way to achieve this kind of visualization - but here is an
> idea: use a scatter line chart and create the dataset so that you have the
> start time and the end time for one projector in different rows. Anyway, it
> won't be easy to achieve this.
> -- Robert
> This posting is provided "AS IS" with no warranties, and confers no rights.
>
> "Nagini Indugula" <NaginiIndugula@.discussions.microsoft.com> wrote in
> message news:DFF5D09D-28B7-4DAB-B86C-5A2EDE25D978@.microsoft.com...
> > Hi:
> >
> > Lets say for eg: you have LCD projectors that you can borrow at various
> > times in a day. I need to create a chart that shows the 24 hours (like
> > 12AM,
> > 1AM, 2Am, 3Am...11PM, 12PM), on the X axis, and the LCD resource name on
> > the
> > Y axis, and if it was booked between 2AM and 4AM, have a colored bar
> > between
> > 2Am and 4Am, and rest blank, and so on for all the resources.
> >
> > What are my options to do something like this. Cant seem to find any way
> > to
> > do this.
> >
> > Thanks
> > NI
>
>