[GitHub] metron issue #873: METRON-1367 Stellar should have some instrumentation of f...

[simonellistonball]

Github user simonellistonball commented on the issue:

https://github.com/apache/metron/pull/873
  
Well, if it helped, cool, let’s keep it... was just curious about your 
use case and whether we should have a bolder scope or if there is anything like 
this that would help elsewhere in general performance debugging. I’m 
certainly not against this in any way. :)

> On 4 Apr 2018, at 22:21, ottobackwards  wrote:
> 
> I'm shelving this. Thinks for taking a look.
> 
> —
> You are receiving this because you were mentioned.
> Reply to this email directly, view it on GitHub, or mute the thread.
> 



---

[GitHub] metron issue #873: METRON-1367 Stellar should have some instrumentation of f...

[ottobackwards]

Github user ottobackwards commented on the issue:

https://github.com/apache/metron/pull/873
  
I'm shelving this.  Thinks for taking a look.



---

[GitHub] metron issue #873: METRON-1367 Stellar should have some instrumentation of f...

[ottobackwards]

Github user ottobackwards commented on the issue:

https://github.com/apache/metron/pull/873
  
@simonellistonball, i have read and replied to you last about 100 times, 
but now I understand.  The PR makes it seem like this is a PR for a major 
instrumentation/sampling solution, when I meant it as a Shell utility to start, 
not a replacement for the existing stellar speed test, but for a different use.

I am sorry for the confusion.  I found this useful in working ON the 
stellar language and functions. But it is not Apache Sirona or anything.  I am 
sorry for the miss understanding.


Thanks





---

[GitHub] metron issue #873: METRON-1367 Stellar should have some instrumentation of f...

[simonellistonball]

Github user simonellistonball commented on the issue:

https://github.com/apache/metron/pull/873
  
Did it work out for you to find bottlenecks @ottobackwards ? I would be 
more inclined to do multi-run micro-benchmarking (I have a bunch of 
multi-sample perfidix runs on all the parser work I've done for example) rather 
than a one of timer. Would a more 'proper' run it a thousand times in the 
proper context of the topology code performance profiler be useful / relevant. 
Did you get the info you need from this as a quicker (and quite possibly good 
enough!) approach?


---

[GitHub] metron issue #873: METRON-1367 Stellar should have some instrumentation of f...

[ottobackwards]

Github user ottobackwards commented on the issue:

https://github.com/apache/metron/pull/873
  
I think if you are writing or testing commands in the repl this would be 
useful.
If it were moved from stellar to common it could be integrated with the 
topologies. 
The reason I submitted it to commons was I thought it might be, but we 
don't have a 'common util' module, long run it should not be in stellar either


---

[GitHub] metron issue #873: METRON-1367 Stellar should have some instrumentation of f...

[nickwallen]

Github user nickwallen commented on the issue:

https://github.com/apache/metron/pull/873
  
IMHO, I just don't see value in capturing timing information like this in 
the REPL.  Timing information like this would be super valuable when run on a 
live cluster in the topologies to help identify bottlenecks.  Is there a way 
that we could capture timings on a live cluster somehow?


---

[GitHub] metron issue #873: METRON-1367 Stellar should have some instrumentation of f...

[ottobackwards]

Github user ottobackwards commented on the issue:

https://github.com/apache/metron/pull/873
  
I think I'm actually going to get to create a new commons- component to 
house this.  Will this pr beat that one?  Who knows?


---