CircuitBreakerTimingTestsOutput_RollingMetricsWithStrategy.txt

Performance with the latest changeset being: 
https://github.com/kristianhald/Polly/commit/7165c90e3dbd5ea75e6048f94e477bf0bcd6f4a5

Iterations per test: 100000
================
Empty delegates only
----------------
Total ticks              : 0000008221
Per 100 iterations ticks : 0000000008
Per iteration ticks      : 0000000000

================
Original CircuitBreaker wrapping empty delegates only
----------------
Total ticks              : 0000081003
Per 100 iterations ticks : 0000000081
Per iteration ticks      : 0000000000

  ================
  Additional from Original CircuitBreaker wrapping empty
  ----------------
  Total ticks              : 0000072782
  Per 100 iterations ticks : 0000000072
  Per iteration ticks      : 0000000000

================
10 good 10 bad, no CircuitBreaker
----------------
Total ticks              : 0002784848
Per 100 iterations ticks : 0000002784
Per iteration ticks      : 0000000027

================
Original CircuitBreaker wrapping 10 good, 10 bad
----------------
Total ticks              : 0005525417
Per 100 iterations ticks : 0000005525
Per iteration ticks      : 0000000055

  ================
  Additional from breaker managing exceptions, and breaking every 10 exceptions
  ----------------
  Total ticks              : 0002740569
  Per 100 iterations ticks : 0000002740
  Per iteration ticks      : 0000000027

================
TimesliceCircuitBreaker wrapping 10 good, 10 bad
----------------
Total ticks              : 0006098737
Per 100 iterations ticks : 0000006098
Per iteration ticks      : 0000000060

  ================
  Additional from TimesliceCircuitBreaker managing exceptions, and breaking every 10 exceptions
  ----------------
  Total ticks              : 0003313889
  Per 100 iterations ticks : 0000003313
  Per iteration ticks      : 0000000033


Circuit breaks during original breaker test: 5000
Circuit breaks during new breaker test     : 5000
Comparability (1 is good): 1

This is the performance for the advanced circuit breaker, where it always uses 10 windows for the entire duration.
The performance program was taken from: https://gist.github.com/reisenberger/92dc8d73b4df127b296ed8daee3ed93d

The difference between this and 'https://gist.github.com/kristianhald/534ed0540e8030060f1c49da04100e4a' is that this uses an interface in the controller, allowing the implementation of the health metrics to be changed without changing the code. Seems like there is about ~10 ticks difference per iteration.