Package io.deephaven.util.grpc

Class FailedChannelDetector<CHANNEL,STATE>

java.lang.Object

io.deephaven.util.grpc.FailedChannelDetector<CHANNEL,STATE>

public class FailedChannelDetector<CHANNEL,STATE>
extends Object

Class to support defining a condition for when a gRPC managed channel is failed, based on either (a) a number of disconnects and reconnects quick in a row, or (b) a sustained period of disconnection. Note that using this class will prevent the channel from staying in IDLE state; if the channel is not configured to do keepalives and avoid IDLE, any time the channel turns to idle using this class will trigger a reconnect attempt on the channel. Note that in some circumstances, the failure callback can be called more than once.

Nested Class Summary

Nested Classes

Modifier and Type	Class	Description
static interface	FailedChannelDetector.ChannelStateReader<CHANNEL,STATE>	Class to avoid a particular import path for gRPC classes, so that different clients can use our functionality for different shadowed gRPC packages.

Constructor Summary

Constructors

Constructor	Description
FailedChannelDetector(String logPfx, ScheduledExecutorService executor, FailedChannelDetector.ChannelStateReader<CHANNEL,STATE> channelStateReader, long channelDownForTimeIsFailurePeriodMillis, int repeatedChannelDownIsFailureCount, long repeatedChannelDownIsFailurePeriodMillis, Runnable failedCb)	Creates the failed channel detector object.

Method Summary

Modifier and Type	Method	Description
void	forceFailure()	Force a failure to be triggered on the next channel state change, no matter what the resulting state is.
void	shutdown()
void	startTracking(CHANNEL channel)	Start tracking for channel failures.
void	stopTracking()	Stop tracking for channel failures

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Details

FailedChannelDetector

public FailedChannelDetector(String logPfx,
 ScheduledExecutorService executor,
 FailedChannelDetector.ChannelStateReader<CHANNEL,STATE> channelStateReader,
 long channelDownForTimeIsFailurePeriodMillis,
 int repeatedChannelDownIsFailureCount,
 long repeatedChannelDownIsFailurePeriodMillis,
 Runnable failedCb)

Creates the failed channel detector object.

Parameters:

logPfx - a prefix String for every log message

executor - a scheduled executor to use for timed checks

channelStateReader - an implementation of stateReader

channelDownForTimeIsFailurePeriodMillis - a period of time where, if the channel stops being ready and does not get back to ready in that period, the channel is considered failed

repeatedChannelDownIsFailureCount - together with the next argument, defines a number of times and a period of time such that, if the channel repeatedly bounces from ready to not ready in that period (or less), the channel will be considered failed.

repeatedChannelDownIsFailurePeriodMillis - together with the previous argument, defines a number of times and a period of time such that, if the channel repeatedly bounces from ready to not ready in that period (or less), the channel will be considered failed.

failedCb - a callback to invoke to notify the a condition for channel failed has been met. Note that in some circumstances, the callback may be invoked more than once.

Method Details

forceFailure

public void forceFailure()

Force a failure to be triggered on the next channel state change, no matter what the resulting state is.

startTracking

public void startTracking(CHANNEL channel)

Start tracking for channel failures. Can only be called once.

Parameters:

channel - the channel to track.

stopTracking

public void stopTracking()

Stop tracking for channel failures

shutdown

public void shutdown()