let me know about new metrics.

540 views
Skip to first unread message

YOTA

unread,
Jan 28, 2013, 11:58:12 PM1/28/13
to storm...@googlegroups.com

Hi all,

I have some questions about new metrics on 0.8.2 as follows,

1.  Is #Executed the number of calling execute() method ?
     I would like to know the definition of it.
2.  How different is "Execute latency" from "Process latency" ?, it includes a period of a tuple waiting on queue. ?
     I would like to know the definition of it, too.
3. is the period of Capacity static ? now it would be last 10m only, i guess. Could I change it?

I could not find the document about the above,
also please let me know the pointer of it.

regards,
 Yoshiki TANAKA.





Nathan Marz

unread,
Jan 30, 2013, 2:50:58 AM1/30/13
to storm...@googlegroups.com
On Mon, Jan 28, 2013 at 8:58 PM, YOTA <yotana...@gmail.com> wrote:

Hi all,

I have some questions about new metrics on 0.8.2 as follows,

1.  Is #Executed the number of calling execute() method ?
     I would like to know the definition of it.

Yes.
 
2.  How different is "Execute latency" from "Process latency" ?, it includes a period of a tuple waiting on queue. ?
     I would like to know the definition of it, too.

Process latency is time until tuple is acked, execute latency is time spent in execute for a tuple
 
3. is the period of Capacity static ? now it would be last 10m only, i guess. Could I change it?

It's static. But it's a derived metric anyway that you can deduce from #executed and execute latency.
 

I could not find the document about the above,
also please let me know the pointer of it.

regards,
 Yoshiki TANAKA.






--
You received this message because you are subscribed to the Google Groups "storm-user" group.
To unsubscribe from this group and stop receiving emails from it, send an email to storm-user+...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

invader zim

unread,
Jan 30, 2013, 4:22:58 AM1/30/13
to storm...@googlegroups.com, nat...@nathanmarz.com
Hi, while we're talking about metrics, is there a way of examining the real-time tuple rate through a bolt/ coming out of a spout (I mean aside from just counting the tuples that go by in a time interval inside your component)

YOTA

unread,
Jan 31, 2013, 9:14:27 PM1/31/13
to storm...@googlegroups.com, nat...@nathanmarz.com

Hi Nathan,

Thank you for all.


2013年1月30日水曜日 16時50分58秒 UTC+9 Nathan Marz:
Reply all
Reply to author
Forward
0 new messages