cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
130651
Views
24
Helpful
17
Replies

how to read cpu utilization

mahesh18
Level 6
Level 6

3433433454843843348742

7798251066103139283961

100

90

80 * * **

70 * * **

60 * * * **

50 * *** * ****

40 ****** ***** *** ****

30 *********************

20 ******************#***

10 ********############**

0....5....1....1....2....2....3....3....4....4....5....5....6....6....7..

0 5 0 5 0 5 0 5 0 5 0 5 0

CPU% per hour (last 72 hours)

* = maximum CPU% # = average CPU%

can some one explain me what is current

cPU and how to read from this graph

thanks

1 Accepted Solution

Accepted Solutions

curci
Level 1
Level 1

When I test my Cisco 1841 router with IOS 12.4T code, the most recent utilization

is on the left of the graph.   I've included two snapshots of the 60-minute graph

displayed with "show proc cpu history".

As an experiment, my 1841 router was lightly utilized, with peak

utilization of 2% to 4% (right side of the graph).  Then I started running a program

to generate traffic to drive up the CPU to 95% about 3 minutes ago (left side of graph).

Obviously, the most current statistics are on the left side of the graph.

I believe that either Rick is mistaken, or Cisco has changed the IOS behavior.

I've included below the appearance of the 60-minute graphs as shown after

3 minutes, 7 minutes, and 27 minutes of high CPU load.

See below,

Ray, CCIE #8037

Graph appearance after running traffic for 3 minutes:

      999

      555242322225212222222222222242222222222222222231221222212242

  100 ***

   90 ##*

   80 ##*

   70 ##*

   60 ##*

   50 ##*

   40 ###

   30 ###

   20 ###

   10 ###        *

     0....5....1....1....2....2....3....3....4....4....5....5....6

               0    5    0    5    0    5    0    5    0    5    0

               CPU% per minute (last 60 minutes)

              * = maximum CPU%   # = average CPU%

      

Here is the same graph as it looks after generating traffic for 7 minutes:

      9999999

      444555524232222521222222222222224222222222222222223122122221

  100    ****

   90 ######*

   80 ######*

   70 ######*

   60 ######*

   50 ######*

   40 #######

   30 #######

   20 #######

   10 #######        *

     0....5....1....1....2....2....3....3....4....4....5....5....6

               0    5    0    5    0    5    0    5    0    5    0

               CPU% per minute (last 60 minutes)

              * = maximum CPU%   # = average CPU%

Here is the same graph as it looks after generating traffic for 27 minutes:

      999999999999999999999999999

      666666666444444444444445555242322225212222222222222242222222

  100 *********              ****

   90 ##########################*

   80 ##########################*

   70 ##########################*

   60 ##########################*

   50 ##########################*

   40 ###########################

   30 ###########################

   20 ###########################

   10 ###########################        *

     0....5....1....1....2....2....3....3....4....4....5....5....6

               0    5    0    5    0    5    0    5    0    5    0

               CPU% per minute (last 60 minutes)

              * = maximum CPU%   # = average CPU%

View solution in original post

17 Replies 17

Joseph W. Doherty
Hall of Fame
Hall of Fame

Not 100% positive, but don't recall whether CPU on most routers can be seen any more current than the average for the last 5 seconds.

The particular graph you've posted, shows the CPU spikes and average, for each hour, going back 72 hours. The left hand column is CPU% , the bottom is each hour (newest on the left), the top is CPU% spike.

Mark Yeates
Level 7
Level 7

The show processes cpu history command displays in ASCII graphical form the total CPU usage on the router over a period of time: one minute, one hour, and 72 hours, displayed in increments of one second, one minute, and one hour, respectively.

Maximum usage is measured and recorded every second; average usage is calculated on periods over one second, But remember that the router may have reached 80 percent only once during that minute, or it may have reached 80 percent multiple times; the router records only the peak reached during the increment and the average over the course of that increment.

If you are worried that your CPU is running too high I would recommend using a more robust monitoring tool to see the actual cpu load in real time. Also remember that the processes can go up during peak traffic patterns.

Mark

Mahesh

Both Joseph and Mark have given good answers about the output of show process cpu history. But neither of them have directly answered the last part of your question. The most current measurement of CPU busy is 10%. As Joseph explained the vertical axis is the measurement of CPU busy and the horizontal axis is the passage of time. So the last entry on the right is the current measurement and it shows 10% busy.

HTH

Rick

HTH

Rick

Hi Rick

so how do u read cpu is now 10 % from this

############ bar?

also do u read reading at told line or bottom horizontal line?

usually we check the avg utilisation (#), but also keep an eye on the max utilisation(*). Bcoz if your device is touching max util for most of the time than you definitely need to think about it. You may want to check which processes are consuming extra resource & review the configs.

Mahesh

The most current measurement in the graph is the entry the the most far to the right. In the graph posted the entry most far to the right had a single "*" and a single entry at that time indicates a CPU busy of 10%.

For any particular time indicated in the graph think of a vertical line at that time location in the graph. Looking at that vertical line you would determine how many symbols (* or #) are in that vertical line, and the number of symbols indicates the amount of CPU busy at that time.

HTH

Rick

HTH

Rick

Rick, correct me if I'm mistaken, but don't all three CPU history graphs show most recent data starting at the left?

Also, although the symbols show peak and average CPU utilization, the numeral values at the top give a more precise peak value. On this chart I see 37% as the extreme left, which I believe is the last hour peak (and 21% as the extreme right value). I suspect the 10% value you note from the graph display of the right most graphic symbols is due to compression of white space.

curci
Level 1
Level 1

View the output with a mono-spaced font like Courier so the columns line up to start. There are 72 integers on the top of the graph written vertically.  The left most column is the most recent utilization 37%.  Reading toward the right, an hour ago, the utilization was 47%. Two hours ago, it was 39%, and so on.  Numbers are shown for the past 3 days or 72 hours. -Ray

Here is the output from a router that has been up just a few hours.  Every hour, the graph shifts one column to the right:

(cpu has been at 15% the past 3 hours)

     

      111612

      555558

  100

   90

   80

   70    *

   60    *

   50    *

   40    *

   30    * *

   20 ******

   10 ******

     0....5....1....1....2....2....3....3....4....4....5....5....6....6....7..

               0    5    0    5    0    5    0    5    0    5    0    5    0

                   CPU% per hour (last 72 hours)

                  * = maximum CPU%   # = average CPU%

Hi all

Have been reading this post...reading CPU on a cisco device has never been easy for me and to be quite honest it still isn't easy.

So hopefully I will be able to understand it now.

Not sure... but it makes more sense to me to have the most recent data on the right.

So if Rick is correct am I right to say that on the graph below the last hour averaged of 10% and spiked around 50% on the same hour?

Hi Everyone,

Seems this post was very helpfull.

Earlier i used to read the graph as opposite to what Rick said.

Thanks to everyone i learned how to read the graph.

Best Regards

Mahesh

curci
Level 1
Level 1

When I test my Cisco 1841 router with IOS 12.4T code, the most recent utilization

is on the left of the graph.   I've included two snapshots of the 60-minute graph

displayed with "show proc cpu history".

As an experiment, my 1841 router was lightly utilized, with peak

utilization of 2% to 4% (right side of the graph).  Then I started running a program

to generate traffic to drive up the CPU to 95% about 3 minutes ago (left side of graph).

Obviously, the most current statistics are on the left side of the graph.

I believe that either Rick is mistaken, or Cisco has changed the IOS behavior.

I've included below the appearance of the 60-minute graphs as shown after

3 minutes, 7 minutes, and 27 minutes of high CPU load.

See below,

Ray, CCIE #8037

Graph appearance after running traffic for 3 minutes:

      999

      555242322225212222222222222242222222222222222231221222212242

  100 ***

   90 ##*

   80 ##*

   70 ##*

   60 ##*

   50 ##*

   40 ###

   30 ###

   20 ###

   10 ###        *

     0....5....1....1....2....2....3....3....4....4....5....5....6

               0    5    0    5    0    5    0    5    0    5    0

               CPU% per minute (last 60 minutes)

              * = maximum CPU%   # = average CPU%

      

Here is the same graph as it looks after generating traffic for 7 minutes:

      9999999

      444555524232222521222222222222224222222222222222223122122221

  100    ****

   90 ######*

   80 ######*

   70 ######*

   60 ######*

   50 ######*

   40 #######

   30 #######

   20 #######

   10 #######        *

     0....5....1....1....2....2....3....3....4....4....5....5....6

               0    5    0    5    0    5    0    5    0    5    0

               CPU% per minute (last 60 minutes)

              * = maximum CPU%   # = average CPU%

Here is the same graph as it looks after generating traffic for 27 minutes:

      999999999999999999999999999

      666666666444444444444445555242322225212222222222222242222222

  100 *********              ****

   90 ##########################*

   80 ##########################*

   70 ##########################*

   60 ##########################*

   50 ##########################*

   40 ###########################

   30 ###########################

   20 ###########################

   10 ###########################        *

     0....5....1....1....2....2....3....3....4....4....5....5....6

               0    5    0    5    0    5    0    5    0    5    0

               CPU% per minute (last 60 minutes)

              * = maximum CPU%   # = average CPU%

Hi Curi,

i also tested seems most recent is on Left hand side.

Lets wait for Ricks reply.

Thanks for replying to the question.

Regards

Mahesh

Great stuff Ray.

Thanks for it.

Hi Ray,

MAny thanks for answering the question.

Now finally i marked this question as answered.

Regards

Mahesh

Getting Started

Find answers to your questions by entering keywords or phrases in the Search bar above. New here? Use these resources to familiarize yourself with the community:

Innovations in Cisco Full Stack Observability - A new webinar from Cisco