PhiAccrualFailureDetector

class PhiAccrualFailureDetector(val threshold: Double, val maxSampleSize: Int, val minStdDeviation: FiniteDuration, val acceptableHeartbeatPause: FiniteDuration, val firstHeartbeatEstimate: FiniteDuration, eventStream: Option[EventStream])(implicit clock: Clock) extends FailureDetector

Implementation of 'The Phi Accrual Failure Detector' by Hayashibara et al. as defined in their paper: [http://www.jaist.ac.jp/~defago/files/pdf/IS_RR_2004_010.pdf]

The suspicion level of failure is given by a value called φ (phi). The basic idea of the φ failure detector is to express the value of φ on a scale that is dynamically adjusted to reflect current network conditions. A configurable threshold is used to decide if φ is considered to be a failure.

The value of φ is calculated as:

φ = -log10(1 - F(timeSinceLastHeartbeat)

where F is the cumulative distribution function of a normal distribution with mean and standard deviation estimated from historical heartbeat inter-arrival times.

Value parameters:
acceptableHeartbeatPause

Duration corresponding to number of potentially lost/delayed heartbeats that will be accepted before considering it to be an anomaly. This margin is important to be able to survive sudden, occasional, pauses in heartbeat arrivals, due to for example garbage collect or network drop.

clock

The clock, returning current time in milliseconds, but can be faked for testing purposes. It is only used for measuring intervals (duration).

firstHeartbeatEstimate

Bootstrap the stats with heartbeats that corresponds to to this duration, with a with rather high standard deviation (since environment is unknown in the beginning)

maxSampleSize

Number of samples to use for calculation of mean and standard deviation of inter-arrival times.

minStdDeviation

Minimum standard deviation to use for the normal distribution used when calculating phi. Too low standard deviation might result in too much sensitivity for sudden, but normal, deviations in heartbeat inter arrival times.

threshold

A low threshold is prone to generate many wrong suspicions but ensures a quick detection in the event of a real crash. Conversely, a high threshold generates fewer mistakes but needs more time to detect actual crashes

Source:
PhiAccrualFailureDetector.scala
class Object
trait Matchable
class Any

Value members

Constructors

def this(threshold: Double, maxSampleSize: Int, minStdDeviation: FiniteDuration, acceptableHeartbeatPause: FiniteDuration, firstHeartbeatEstimate: FiniteDuration)(implicit clock: Clock)

Constructor without eventStream to support backwards compatibility

Constructor without eventStream to support backwards compatibility

Source:
PhiAccrualFailureDetector.scala
def this(config: Config, ev: EventStream)

Constructor that reads parameters from config. Expecting config properties named threshold, max-sample-size, min-std-deviation, acceptable-heartbeat-pause and heartbeat-interval.

Constructor that reads parameters from config. Expecting config properties named threshold, max-sample-size, min-std-deviation, acceptable-heartbeat-pause and heartbeat-interval.

Source:
PhiAccrualFailureDetector.scala

Concrete methods

final override def heartbeat(): Unit
def phi: Double

The suspicion level of the accrual failure detector.

The suspicion level of the accrual failure detector.

If a connection does not have any records in failure detector then it is considered healthy.

Source:
PhiAccrualFailureDetector.scala