| Author |
Message
|
| fedux |
 Posted: Mon May 16, 2011 9:36 am Post subject: time spent by a message |
 |
|
Novice
Joined: 12 Jan 2007
Posts: 20
|
hi,
I need to measure the time between when a message enters the qmanager until it exits. Particularly I need to know the time a message is placed in a remote queue until the other qmanager gives me the ok.
The part of statistics and accountig do not give me the precision I need.
The trace seems to be the most appropriate tool, but it is difficult to understand.
could someone help me ?
Thanks |
|
| Back to top |
|
 |
| bruce2359 |
| Posted: Mon May 16, 2011 10:21 am Post subject: |
|
|
Poobah
Joined: 05 Jan 2008
Posts: 9469
Location: US: west coast, almost. Otherwise, enroute.
|
You want to do this for one particular message? Or all messages?
_________________
I like deadlines. I like to wave as they pass by.
ב''ה
Lex Orandi, Lex Credendi, Lex Vivendi. As we Worship, So we Believe, So we Live. |
|
| Back to top |
|
|
| fedux |
| Posted: Mon May 16, 2011 10:42 am Post subject: |
|
|
Novice
Joined: 12 Jan 2007
Posts: 20
|
There are 7 messages that are generated to build the home site index
thanks |
|
| Back to top |
|
|
| bruce2359 |
| Posted: Mon May 16, 2011 10:55 am Post subject: |
|
|
Poobah
Joined: 05 Jan 2008
Posts: 9469
Location: US: west coast, almost. Otherwise, enroute.
|
| fedux wrote: |
| There are 7 messages that are generated to build the home site index thanks |
I don't understand that you are telling me. Is this an answer to my questions?
_________________
I like deadlines. I like to wave as they pass by.
ב''ה
Lex Orandi, Lex Credendi, Lex Vivendi. As we Worship, So we Believe, So we Live. |
|
| Back to top |
|
|
| fedux |
| Posted: Mon May 16, 2011 11:03 am Post subject: |
|
|
Novice
Joined: 12 Jan 2007
Posts: 20
|
sory, yes for a particular message
thanks |
|
| Back to top |
|
|
| bruce2359 |
| Posted: Mon May 16, 2011 11:08 am Post subject: Re: time spent by a message |
|
|
Poobah
Joined: 05 Jan 2008
Posts: 9469
Location: US: west coast, almost. Otherwise, enroute.
|
| fedux wrote: |
| ... until the other qmanager gives me the ok. |
What exactly do you mean by this?
_________________
I like deadlines. I like to wave as they pass by.
ב''ה
Lex Orandi, Lex Credendi, Lex Vivendi. As we Worship, So we Believe, So we Live. |
|
| Back to top |
|
|
| bruce2359 |
| Posted: Mon May 16, 2011 11:20 am Post subject: |
|
|
Poobah
Joined: 05 Jan 2008
Posts: 9469
Location: US: west coast, almost. Otherwise, enroute.
|
Take a look at the dspmqrte control command.
_________________
I like deadlines. I like to wave as they pass by.
ב''ה
Lex Orandi, Lex Credendi, Lex Vivendi. As we Worship, So we Believe, So we Live. |
|
| Back to top |
|
|
| fedux |
| Posted: Mon May 16, 2011 1:11 pm Post subject: |
|
|
Novice
Joined: 12 Jan 2007
Posts: 20
|
ok, I'm going to simplify, I need to know how long it takes the manager to accept a message and make the delivery to another manager
thanks |
|
| Back to top |
|
|
| mvic |
| Posted: Mon May 16, 2011 1:28 pm Post subject: |
|
|
Jedi
Joined: 09 Mar 2004
Posts: 2080
|
| fedux wrote: |
| ok, I'm going to simplify, I need to know how long it takes the manager to accept a message and make the delivery to another manager |
Assuming your channels are ok, and everything else is ok, this should be no more than a few milliseconds. It'll be longer if there are problems anywhere.
I'm wondering, what calculations were you hoping to do with the times you measured?
There is a time-on-queue statistic somewhere I think, hopefully can be found in the documentation. Search for QTIME. |
|
| Back to top |
|
|
| fjb_saper |
| Posted: Mon May 16, 2011 7:47 pm Post subject: |
|
|
Grand High Poobah
Joined: 18 Nov 2003
Posts: 20756
Location: LI,NY
|
Additionally if you set qmonitoring to low your monitoring software will pick up the age of the oldest message on queue. (in seconds)
Note that factored in there will be transit times, plus time waiting on queue for the consumer (including any delaying due to priority delivery)
Have fun 
_________________
MQ & Broker admin |
|
| Back to top |
|
|
| fedux |
| Posted: Tue May 17, 2011 5:12 am Post subject: |
|
|
Novice
Joined: 12 Jan 2007
Posts: 20
|
I know, I know it is tiny, but I need to know how many milliseconds consumed!!
I have to prove that mq is not what is consuming the time ..!!!!!!
thanks |
|
| Back to top |
|
|
| mqjeff |
| Posted: Tue May 17, 2011 5:16 am Post subject: |
|
|
Grand Master
Joined: 25 Jun 2008
Posts: 17447
|
| dspmqrte will show you how long it takes similar messages to traverse the same route. |
|
| Back to top |
|
|
| exerk |
| Posted: Tue May 17, 2011 5:43 am Post subject: |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
You might also want to think about monitoring the queue service interval - the time spent traversing queue managers may be milliseconds but if there is a large gap between when the MCA PUTs the message(s) on the queue, and when the application GETs the message(s) from the queue, MQ will be perceived to be the problem (false positive). In other words, your transit time may be negligible but the time it takes for an application to remove the message may not.
_________________
It's puzzling, I don't think I've ever seen anything quite like this before...and it's hard to soar like an eagle when you're surrounded by turkeys. |
|
| Back to top |
|
|
| fedux |
| Posted: Tue May 17, 2011 5:54 am Post subject: |
|
|
Novice
Joined: 12 Jan 2007
Posts: 20
|
ok, we already tried it but I need to measure the application message, there is some way?
thanks |
|
| Back to top |
|
|
| exerk |
| Posted: Tue May 17, 2011 6:08 am Post subject: |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
Please clarify the below statement:
| fedux wrote: |
| ...I need to measure the application message... |
It seems to me that you need to de-construct the elements of the end-to-end track of the message, with particular regard to:
1. Time taken for the originating application to form the message(s), PUT and commit*;
2. Transit time of message between queue managers;
3. Time taken for the consuming application to GET the message(s), process the content and commit*.
The above are always going to be variable anyway, but just what is the issue here? Whom is so concerned with transit times? Why are they concerned?
* commit is not used in the sense of UoW, merely to describe the end of the PUT/GET process
_________________
It's puzzling, I don't think I've ever seen anything quite like this before...and it's hard to soar like an eagle when you're surrounded by turkeys. |
|
| Back to top |
|
|
|
|
