Hystrix Configuration

Question

I am trying to implement hystrix for my application using hystrix-javanica.

I have configured hystrix-configuration.properties as below

hystrix.command.default.execution.isolation.strategy=SEMAPHORE
hystrix.command.default.execution.isolation.thread.timeoutInMilliseconds=10000 
hystrix.command.default.fallback.enabled=true
hystrix.command.default.circuitBreaker.enabled=true
hystrix.command.default.circuitBreaker.requestVolumeThreshold=3 
hystrix.command.default.circuitBreaker.sleepWindowInMilliseconds=50000
hystrix.command.default.circuitBreaker.errorThresholdPercentage=50

short-circuit pattern is working fine but i have a doubt in this hystrix.command.default.circuitBreaker.requestVolumeThreshold=3

1. Is it stating open the circuit after 3 failures or
2. Open the circuit after 3 concurrent failures.

Gone through the documentation link

Can anybody answer?

 

 

Answer 1

How Hystrix Circuit-Breaker operates: Hystrix does not offer a circuit breaker which breaks after a given number of failures. The Hystrix circuit will break if:

within a timespan of duration metrics.rollingStats.timeInMilliseconds, the percentage of actions resulting in a handled exception exceeds errorThresholdPercentage, provided also that the number of actions through the circuit in the timespan is at least requestVolumeThreshold

---

What is requestVolumeThreshold? requestVolumeThreshold is a minimum threshold for the volume (number) of calls through the circuit that must be met (within the rolling window), before the circuit calculates a percentage failure rate at all. Only when this minimum volume (in each time window) has been met, will the circuit compare the failure proportion of your calls against the errorThresholdPercentage you have configured.

Imagine there was no such minimum-volume-through-the-circuit threshold. Imagine the first call in a time window errors. You would have 1 of 1 calls being an error, = 100% failure rate, which is higher than the 50% threshold you have set. So the circuit would break immediately.

The requestVolumeThreshold exists so that this does not happen. It's effectively saying, the error rate through your circuit isn't statistically significant (and won't be compared against errorThresholdPercentage) until at least requestVolumeThreshold calls have been received in each time window.