Package com.netflix.astyanax.connectionpool.impl

Interface Summary

HostStats
SimpleHostConnectionPool.Listener<CL>	Interface to notify the owning connection pool of up/down state changes.
Topology<CL>

Class Summary

AbstractExecuteWithFailoverImpl<CL,R>	Abstract class that provides a basic impl of ExecuteWithFailover.
AbstractExecutionImpl<R>	Abstract impl that repeatedly executes while consulting a RetryPolicy
AbstractHostPartitionConnectionPool<CL>	Base for all connection pools that keep a separate pool of connections for each host.
AbstractLatencyScoreStrategyImpl
AbstractOperationFilter<CL,R>	Class that wraps an Operation to provide extra functionality.
AbstractTopology<CL>
BadHostDetectorImpl	BadHostDetector which marks the host as failed if there is a transport exception or if it timed out too many times within a certain time window
BagOfConnectionsConnectionPoolImpl<CL>	Connection pool which puts all connections in a single queue.
ConnectionPoolConfigurationImpl	Basic impl for ConnectionPoolConfiguration that uses a bunch of defaults for al the connection pool config.
ConnectionPoolMBeanManager	Simple jmx bean manager.
CountingConnectionPoolMonitor	Impl for ConnectionPoolMonitor that employs counters to track stats such as operation success / failures / timeouts / socket timeouts / interrupted connection created / borrowed / returned / closed / create failures hosts added / removed / marked as down / reactivated transport failures and other useful stats
EmaLatencyScoreStrategyImpl	Calculate latency as an exponential moving average.
ExponentialRetryBackoffStrategy	Impl for RetryBackoffStrategy that is used to reconnect a HostConnectionPool when a host is marked as down.
FixedRetryBackoffStrategy	Impl for RetryBackoffStrategy that is used to reconnect a HostConnectionPool when a host is marked as down.
HostConnectionPoolPartition<CL>	Collection of hosts that are grouped by a certain criteria (such as token or rack) The class maintains the list of active pools for the given criteria (such as a token range in a ring) Any changes to the pools in this partition causes the active set to be revised and sorted according to the latency score strategy provided.
LeastOutstandingExecuteWithFailover<CL,R>	Finds the HostConnectionPool with the max idle connections when borrowing a connection from the pool.
NodeDiscoveryImpl	Re-discover the ring on a fixed interval to identify new nodes or changes to the ring tokens.
NodeDiscoveryMonitorManager	Jmx bean monitor manager for NodeDiscoveryMonitor
OldHostSupplierAdapter
OperationFilterFactoryList	Uses the decorator pattern to maintain a list of OperationFilterFactory for the specified Operation
OperationResultImpl<R>	Impl for OperationResult Tracks operation attempts, operation pinned host and latency associated with the operation.
RoundRobinConnectionPoolImpl<CL>	Connection pool implementation using simple round robin.
RoundRobinExecuteWithFailover<CL,R>	Class that extends AbstractExecuteWithFailoverImpl to provide functionality for borrowing a Connection from a list of HostConnectionPool(s) in a round robin fashion.
SimpleAuthenticationCredentials
SimpleHostConnectionPool<CL>	Pool of connections for a single host and implements the HostConnectionPool interface
SimpleRateLimiterImpl
Slf4jConnectionPoolMonitorImpl	Impl for CountingConnectionPoolMonitor that does not track any internal state / counters.
SmaLatencyScoreStrategyImpl
TokenAwareConnectionPoolImpl<CL>	Connection pool that partitions connections by the hosts which own the token being operated on.
TokenHostConnectionPoolPartition<CL>	Collection of pools that own a token range of the ring
TokenPartitionedTopology<CL>	Partition hosts by start token.
TokenRangeImpl	Impl for TokenRange

Enum Summary

ConnectionPoolType	This is referenced from the AstyanaxContext.Builder class for instantiating the right ConnectionPool
HostSelectorStrategy	See AbstractHostPartitionConnectionPool.executeWithFailover(com.netflix.astyanax.connectionpool.Operation, com.netflix.astyanax.retry.RetryPolicy) for details on how the strategy is consulted when executing an operation with a specific operation failover strategy.