| Author |
Message
|
| sleepyjamie |
 Posted: Tue Sep 01, 2015 8:35 am Post subject: Fetching milliseconds from CURRENT_TIMESTAMP |
 |
|
Centurion
Joined: 29 Apr 2015
Posts: 135
|
I have a message flow that I want to measure the http response time which will be sent to InfluxDB/Graphana.
| Code: |
| Compute Node->HTTP Request->Trace Node |
In my compute node I build the http payload and also
| Code: |
| SET Environment.HTTP_REQUEST_TIMESTAMP = CURRENT_TIMESTAMP |
In my trace node I have the following in the pattern
| Code: |
| measurement http_response_time=${CAST ((Environment.HTTP_REQUEST_TIMESTAMP - CURRENT_TIMESTAMP) SECOND as INTEGER)} |
Which outputs the follow to my log file:
| Code: |
| measurement http_response_time=0 |
What I would like is that instead of outputting SECONDS I want it to output milliseconds. For the life of me I cannot find documentation on how to do this. |
|
| Back to top |
|
 |
| smdavies99 |
| Posted: Tue Sep 01, 2015 8:48 am Post subject: |
|
|
Jedi Council
Joined: 10 Feb 2003
Posts: 6076
Location: Somewhere over the Rainbow this side of Never-never land.
|
Think about the CAST.
Lets for argument sake say that the difference in time between the two times is 1.2345 seconds.
Casting that to an integer will give 1
So your value of '0' is probably correct if the elapsed time is less than one second.
So to get the milliseconds I might do something like this.
1) take the time difference in SECONDS and cast to FLOAT
2) Multiply by 1000
3) Cast to INTEGER
so for
1) 0.123456
2) 0.1234 * 1000 = 1234.56
3) Cast to Integer = 1234 milliseconds
I hope this gives you an idea of how to go forward.
[/code]
_________________
WMQ User since 1999
MQSI/WBI/WMB/'Thingy' User since 2002
Linux user since 1995
Every time you reinvent the wheel the more square it gets (anon). If in doubt think and investigate before you ask silly questions. |
|
| Back to top |
|
|
| sleepyjamie |
| Posted: Tue Sep 01, 2015 8:53 am Post subject: |
|
|
Centurion
Joined: 29 Apr 2015
Posts: 135
|
Yeah just realized the alternative is to cast as float and multiply by 1000.
I was hoping to extract it as milliseconds however that doesn't work.
6/12
1/2
thanks! |
|
| Back to top |
|
|
| smdavies99 |
| Posted: Tue Sep 01, 2015 9:10 am Post subject: |
|
|
Jedi Council
Joined: 10 Feb 2003
Posts: 6076
Location: Somewhere over the Rainbow this side of Never-never land.
|
You may have to use the MOD function
{from memory}
Personaly, I'd do this in ESQL and then print out the result in the trace node.
At least then you can debug the logic.
Yay! Post Number 5000.
_________________
WMQ User since 1999
MQSI/WBI/WMB/'Thingy' User since 2002
Linux user since 1995
Every time you reinvent the wheel the more square it gets (anon). If in doubt think and investigate before you ask silly questions. |
|
| Back to top |
|
|
| Vitor |
| Posted: Tue Sep 01, 2015 9:22 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| smdavies99 wrote: |
| Yay! Post Number 5000. |
Welcome To The Council
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| smdavies99 |
| Posted: Tue Sep 01, 2015 9:47 am Post subject: |
|
|
Jedi Council
Joined: 10 Feb 2003
Posts: 6076
Location: Somewhere over the Rainbow this side of Never-never land.
|
| Vitor wrote: |
| smdavies99 wrote: |
| Yay! Post Number 5000. |
Welcome To The Council |
Thank you kind sir.
_________________
WMQ User since 1999
MQSI/WBI/WMB/'Thingy' User since 2002
Linux user since 1995
Every time you reinvent the wheel the more square it gets (anon). If in doubt think and investigate before you ask silly questions. |
|
| Back to top |
|
|
| maurito |
| Posted: Wed Sep 02, 2015 1:51 am Post subject: |
|
|
Partisan
Joined: 17 Apr 2014
Posts: 358
|
| sleepyjamie wrote: |
Yeah just realized the alternative is to cast as float and multiply by 1000.
I was hoping to extract it as milliseconds however that doesn't work.
6/12
1/2
thanks! |
save the Environment.HTTP_REQUEST_TIMESTAMP as ss.SSSSS , for example:
| Code: |
SET Environment.HTTP_REQUEST_TIMESTAMP = SUBSTRING(CAST(CURRENT_TIMESTAMP AS CHAR FORMAT 'yyyy-MM-dd HH:mm:ss.SSSSSS')
FROM LENGTH('yyyy-MM-dd HH:mm:.')); |
and cast as decimal.
do the same in the trace node, then substract the two values ( in your example above the substruction should be the other way around ! , as the time in the trace node will be greater that the one in the Environment ).
mind you, that is not going to work either... need to consider the case when you change from seconds to minutes, minutes to hours... ignore the suggestion.
so probably you need to do MINUTE TO SECOND, or SECOND, which you are already doing. |
|
| Back to top |
|
|
| smdavies99 |
| Posted: Wed Sep 02, 2015 2:57 am Post subject: |
|
|
Jedi Council
Joined: 10 Feb 2003
Posts: 6076
Location: Somewhere over the Rainbow this side of Never-never land.
|
Here is how I did it in an old project.
In a compute node
| Code: |
set Environment.DATA.CurTime_1 = CURRENT_TIMESTAMP;
declare fDiff FLOAT;
set fDiff = CAST((Environment.DATA.CurTime_1 - Environment.DATA.CurTime_0) SECOND as FLOAT);
set fDiff = fDiff * 1000.0;
declare iTime INTEGER;
set iTime = CAST(fDiff as integer);
set OutputRoot.XMLNSC.Data.T1 = Environment.DATA.CurTime_0;
set OutputRoot.XMLNSC.Data.T2 = Environment.DATA.CurTime_1;
set OutputRoot.XMLNSC.Data.D1 = fDiff;
set OutputRoot.XMLNSC.Data.I1 = iTime;
|
Then in the Trace node
| Code: |
Elapsed time = ${Root.XMLNSC.Data.I1} Milliseconds
|
With the input data of
| Code: |
CurTime_0 = TIMESTAMP '2010-04-12 18:52:44.468712' (TIMESTAMP)
CurTime_1 = TIMESTAMP '2010-04-12 18:52:44.854580' (TIMESTAMP)
|
I got this in the output
| Code: |
Elapsed time = 385 Milliseconds
|
Enjoy
_________________
WMQ User since 1999
MQSI/WBI/WMB/'Thingy' User since 2002
Linux user since 1995
Every time you reinvent the wheel the more square it gets (anon). If in doubt think and investigate before you ask silly questions. |
|
| Back to top |
|
|
| sleepyjamie |
| Posted: Wed Sep 02, 2015 7:57 am Post subject: |
|
|
Centurion
Joined: 29 Apr 2015
Posts: 135
|
Here is the solution I ended up going with. I didn't want a lot of boilerplate code in compute nodes. Instead I did all the logic to determine milliseconds in the Trace Node.
| Code: |
| measurement http_response_time=${CAST(CAST ((CURRENT_TIMESTAMP - Environment.GCM_HTTP_REQUEST_TIMESTAMP) SECOND as DECIMAL) * 1000 AS INTEGER)}} |
|
|
| Back to top |
|
|
| smdavies99 |
| Posted: Wed Sep 02, 2015 8:08 am Post subject: |
|
|
Jedi Council
Joined: 10 Feb 2003
Posts: 6076
Location: Somewhere over the Rainbow this side of Never-never land.
|
You could have put it all in a re-usable subflow. That's where mine is. Ready to use whenever needed.
Ok, I've not had to use it for a couple of years but... You never know when the need would arise again.
_________________
WMQ User since 1999
MQSI/WBI/WMB/'Thingy' User since 2002
Linux user since 1995
Every time you reinvent the wheel the more square it gets (anon). If in doubt think and investigate before you ask silly questions. |
|
| Back to top |
|
|
| sleepyjamie |
| Posted: Wed Sep 02, 2015 8:52 am Post subject: |
|
|
Centurion
Joined: 29 Apr 2015
Posts: 135
|
| smdavies99 wrote: |
You could have put it all in a re-usable subflow. That's where mine is. Ready to use whenever needed.
Ok, I've not had to use it for a couple of years but... You never know when the need would arise again. |
Yes that's another option. Thanks! |
|
| Back to top |
|
|
|
|
