Re: Review Request 26376: Implementing non-prod MTTA/R SLA metrics.

[Maxim Khutornenko]

---
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/26376/
---

(Updated Oct. 9, 2014, 9:04 p.m.)


Review request for Aurora and Bill Farner.


Changes
---

CR comments.


Bugs: AURORA-774
https://issues.apache.org/jira/browse/AURORA-774


Repository: aurora


Description
---

Implementing non-prod MTTA/R SLA metrics.


Diffs (updated)
-

  src/main/java/org/apache/aurora/scheduler/sla/MetricCalculator.java 
dca680415426be2bc760875d5774c9e9399ea94b 
  src/main/java/org/apache/aurora/scheduler/sla/SlaAlgorithm.java 
33b6bbe4b39d516595276128bbfa0686d0bb9cbd 
  src/test/java/org/apache/aurora/scheduler/sla/MetricCalculatorTest.java 
aeb90bbb822254cbe4691e45092b9581596ad800 
  src/test/java/org/apache/aurora/scheduler/sla/SlaTestUtil.java 
96a04389e6b14c4d1c0bdb4d06abc046e7ea2151 

Diff: https://reviews.apache.org/r/26376/diff/


Testing
---

./gradlew -Pq build


Thanks,

Maxim Khutornenko

Re: Review Request 26376: Implementing non-prod MTTA/R SLA metrics.

[Maxim Khutornenko]

> On Oct. 9, 2014, 2:46 p.m., Bill Farner wrote:
> > src/test/java/org/apache/aurora/scheduler/sla/MetricCalculatorTest.java, 
> > line 59
> > 
> >
> > Can you instead export these values into a Map and 
> > validate the full map at the end?  That would firm up the value checking 
> > here.
> > 
> > Something like:
> > 
> > expect(...makeGauge(..)).andAnswer(
> > new IAnswer<..>() {
> >   ...
> > AtomicLong value = new AtomicLong().
> > map.put(key, value).
> > return new StatImpl()...;
> > }

Thought about that but it would require hardcoding all 27 of currently exposed 
metrics along with their values. I have modified a bit to fully validate metric 
names instead. That's the only validation suitable for this test. Values are 
asserted in SlaAlgorithmTest.


- Maxim


---
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/26376/#review55984
---


On Oct. 6, 2014, 7:10 p.m., Maxim Khutornenko wrote:
> 
> ---
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/26376/
> ---
> 
> (Updated Oct. 6, 2014, 7:10 p.m.)
> 
> 
> Review request for Aurora and Bill Farner.
> 
> 
> Bugs: AURORA-774
> https://issues.apache.org/jira/browse/AURORA-774
> 
> 
> Repository: aurora
> 
> 
> Description
> ---
> 
> Implementing non-prod MTTA/R SLA metrics.
> 
> 
> Diffs
> -
> 
>   src/main/java/org/apache/aurora/scheduler/sla/MetricCalculator.java 
> dca680415426be2bc760875d5774c9e9399ea94b 
>   src/main/java/org/apache/aurora/scheduler/sla/SlaAlgorithm.java 
> 33b6bbe4b39d516595276128bbfa0686d0bb9cbd 
>   src/test/java/org/apache/aurora/scheduler/sla/MetricCalculatorTest.java 
> aeb90bbb822254cbe4691e45092b9581596ad800 
>   src/test/java/org/apache/aurora/scheduler/sla/SlaTestUtil.java 
> 96a04389e6b14c4d1c0bdb4d06abc046e7ea2151 
> 
> Diff: https://reviews.apache.org/r/26376/diff/
> 
> 
> Testing
> ---
> 
> ./gradlew -Pq build
> 
> 
> Thanks,
> 
> Maxim Khutornenko
> 
>

Re: Review Request 26376: Implementing non-prod MTTA/R SLA metrics.

[Bill Farner]

---
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/26376/#review55984
---

Ship it!



src/test/java/org/apache/aurora/scheduler/sla/MetricCalculatorTest.java


Can you instead export these values into a Map and 
validate the full map at the end?  That would firm up the value checking here.

Something like:

expect(...makeGauge(..)).andAnswer(
new IAnswer<..>() {
  ...
AtomicLong value = new AtomicLong().
map.put(key, value).
return new StatImpl()...;
}


- Bill Farner


On Oct. 6, 2014, 7:10 p.m., Maxim Khutornenko wrote:
> 
> ---
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/26376/
> ---
> 
> (Updated Oct. 6, 2014, 7:10 p.m.)
> 
> 
> Review request for Aurora and Bill Farner.
> 
> 
> Bugs: AURORA-774
> https://issues.apache.org/jira/browse/AURORA-774
> 
> 
> Repository: aurora
> 
> 
> Description
> ---
> 
> Implementing non-prod MTTA/R SLA metrics.
> 
> 
> Diffs
> -
> 
>   src/main/java/org/apache/aurora/scheduler/sla/MetricCalculator.java 
> dca680415426be2bc760875d5774c9e9399ea94b 
>   src/main/java/org/apache/aurora/scheduler/sla/SlaAlgorithm.java 
> 33b6bbe4b39d516595276128bbfa0686d0bb9cbd 
>   src/test/java/org/apache/aurora/scheduler/sla/MetricCalculatorTest.java 
> aeb90bbb822254cbe4691e45092b9581596ad800 
>   src/test/java/org/apache/aurora/scheduler/sla/SlaTestUtil.java 
> 96a04389e6b14c4d1c0bdb4d06abc046e7ea2151 
> 
> Diff: https://reviews.apache.org/r/26376/diff/
> 
> 
> Testing
> ---
> 
> ./gradlew -Pq build
> 
> 
> Thanks,
> 
> Maxim Khutornenko
> 
>

Re: Review Request 26376: Implementing non-prod MTTA/R SLA metrics.

[Maxim Khutornenko]

> On Oct. 8, 2014, 5:32 p.m., Bill Farner wrote:
> > src/main/java/org/apache/aurora/scheduler/sla/MetricCalculator.java, line 81
> > 
> >
> > Seems like we're moving towards identical calculations for prod and 
> > non-prod tasks.  Should we just do that now and avoid the independent 
> > groupings?

I don't necessarily see prods and non-prods converging any time soon. Doing 
that now would require more work to exclude non-prods from platform uptime 
calculations and would add redundant job uptime metrics. I'd rather keep them 
separate at least for now and see if it ever becomes a trend.


- Maxim


---
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/26376/#review55829
---


On Oct. 6, 2014, 7:10 p.m., Maxim Khutornenko wrote:
> 
> ---
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/26376/
> ---
> 
> (Updated Oct. 6, 2014, 7:10 p.m.)
> 
> 
> Review request for Aurora and Bill Farner.
> 
> 
> Bugs: AURORA-774
> https://issues.apache.org/jira/browse/AURORA-774
> 
> 
> Repository: aurora
> 
> 
> Description
> ---
> 
> Implementing non-prod MTTA/R SLA metrics.
> 
> 
> Diffs
> -
> 
>   src/main/java/org/apache/aurora/scheduler/sla/MetricCalculator.java 
> dca680415426be2bc760875d5774c9e9399ea94b 
>   src/main/java/org/apache/aurora/scheduler/sla/SlaAlgorithm.java 
> 33b6bbe4b39d516595276128bbfa0686d0bb9cbd 
>   src/test/java/org/apache/aurora/scheduler/sla/MetricCalculatorTest.java 
> aeb90bbb822254cbe4691e45092b9581596ad800 
>   src/test/java/org/apache/aurora/scheduler/sla/SlaTestUtil.java 
> 96a04389e6b14c4d1c0bdb4d06abc046e7ea2151 
> 
> Diff: https://reviews.apache.org/r/26376/diff/
> 
> 
> Testing
> ---
> 
> ./gradlew -Pq build
> 
> 
> Thanks,
> 
> Maxim Khutornenko
> 
>

Re: Review Request 26376: Implementing non-prod MTTA/R SLA metrics.

[Bill Farner]

---
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/26376/#review55829
---



src/main/java/org/apache/aurora/scheduler/sla/MetricCalculator.java


Seems like we're moving towards identical calculations for prod and 
non-prod tasks.  Should we just do that now and avoid the independent groupings?


- Bill Farner


On Oct. 6, 2014, 7:10 p.m., Maxim Khutornenko wrote:
> 
> ---
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/26376/
> ---
> 
> (Updated Oct. 6, 2014, 7:10 p.m.)
> 
> 
> Review request for Aurora and Bill Farner.
> 
> 
> Bugs: AURORA-774
> https://issues.apache.org/jira/browse/AURORA-774
> 
> 
> Repository: aurora
> 
> 
> Description
> ---
> 
> Implementing non-prod MTTA/R SLA metrics.
> 
> 
> Diffs
> -
> 
>   src/main/java/org/apache/aurora/scheduler/sla/MetricCalculator.java 
> dca680415426be2bc760875d5774c9e9399ea94b 
>   src/main/java/org/apache/aurora/scheduler/sla/SlaAlgorithm.java 
> 33b6bbe4b39d516595276128bbfa0686d0bb9cbd 
>   src/test/java/org/apache/aurora/scheduler/sla/MetricCalculatorTest.java 
> aeb90bbb822254cbe4691e45092b9581596ad800 
>   src/test/java/org/apache/aurora/scheduler/sla/SlaTestUtil.java 
> 96a04389e6b14c4d1c0bdb4d06abc046e7ea2151 
> 
> Diff: https://reviews.apache.org/r/26376/diff/
> 
> 
> Testing
> ---
> 
> ./gradlew -Pq build
> 
> 
> Thanks,
> 
> Maxim Khutornenko
> 
>

Re: Review Request 26376: Implementing non-prod MTTA/R SLA metrics.

[Maxim Khutornenko]

---
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/26376/
---

(Updated Oct. 6, 2014, 7:10 p.m.)


Review request for Aurora and Bill Farner.


Bugs: AURORA-774
https://issues.apache.org/jira/browse/AURORA-774


Repository: aurora


Description
---

Implementing non-prod MTTA/R SLA metrics.


Diffs
-

  src/main/java/org/apache/aurora/scheduler/sla/MetricCalculator.java 
dca680415426be2bc760875d5774c9e9399ea94b 
  src/main/java/org/apache/aurora/scheduler/sla/SlaAlgorithm.java 
33b6bbe4b39d516595276128bbfa0686d0bb9cbd 
  src/test/java/org/apache/aurora/scheduler/sla/MetricCalculatorTest.java 
aeb90bbb822254cbe4691e45092b9581596ad800 
  src/test/java/org/apache/aurora/scheduler/sla/SlaTestUtil.java 
96a04389e6b14c4d1c0bdb4d06abc046e7ea2151 

Diff: https://reviews.apache.org/r/26376/diff/


Testing
---

./gradlew -Pq build


Thanks,

Maxim Khutornenko