Jump to: navigation, search

Difference between revisions of "Ceilometer/AlarmEvaluatorJuno"

(Created page with "====Statistics over sliding windows==== Statistics where increments/decrements are possible: * count, * sum, * average, * stddev Statistics with special implementation: *...")
 
(Statistics over sliding windows)
 
(4 intermediate revisions by the same user not shown)
Line 1: Line 1:
====Statistics over sliding windows====
+
==Statistics over sliding windows==
 +
Two approaches to improve the alarm evaluator were discussed at the summit.
 +
 
 +
===Map-reduce approach===
 +
[[File:EvaluatorMapReduce.png]]
 +
<br/><br/><br/><br/>
 +
 
 +
===Increment/decrement approach===
 +
[[File:EvaluatorIncDecrement.png]]
 +
<br/>
 +
 
 
Statistics where increments/decrements are possible:  
 
Statistics where increments/decrements are possible:  
 
* count,  
 
* count,  
Line 5: Line 15:
 
* average,  
 
* average,  
 
* stddev
 
* stddev
 +
 
Statistics with special implementation:
 
Statistics with special implementation:
 
* min/max : needs O(log n) space (probably in DB), where n = window size
 
* min/max : needs O(log n) space (probably in DB), where n = window size
 
* percentile : classical statistics query
 
* percentile : classical statistics query
 
* cardinality : approximate algorithm, extremely efficient
 
* cardinality : approximate algorithm, extremely efficient
 +
 +
Map-reduce approach covers all these statistics.

Latest revision as of 19:11, 16 May 2014

Statistics over sliding windows

Two approaches to improve the alarm evaluator were discussed at the summit.

Map-reduce approach

EvaluatorMapReduce.png



Increment/decrement approach

EvaluatorIncDecrement.png

Statistics where increments/decrements are possible:

  • count,
  • sum,
  • average,
  • stddev

Statistics with special implementation:

  • min/max : needs O(log n) space (probably in DB), where n = window size
  • percentile : classical statistics query
  • cardinality : approximate algorithm, extremely efficient

Map-reduce approach covers all these statistics.