Package com.google.cloud.bigtable.data.v2

Class BigtableDataSettings.Builder

  • Enclosing class:
    BigtableDataSettings
    public static class BigtableDataSettings.Builder
extends Object
    Builder for BigtableDataSettings.

    • Method Detail

      • getProjectId

        public String getProjectId()
        Gets the project id that was previously set on this Builder.

      • getInstanceId

        public String getInstanceId()
        Gets the instance id that was previously set on this Builder.

      • setAppProfileId

        public BigtableDataSettings.Builder setAppProfileId​(@Nonnull
                                                    String appProfileId)
        Sets the AppProfile to use.

        An application profile (sometimes also shortened to "app profile") is a group of configuration parameters for an individual use case. A client will identify itself with an application profile ID at connection time, and the requests will be handled according to that application profile.

      • setDefaultAppProfileId

        public BigtableDataSettings.Builder setDefaultAppProfileId()
        Resets the AppProfile id to the default for the instance.

        An application profile (sometimes also shortened to "app profile") is a group of configuration parameters for an individual use case. A client will identify itself with an application profile ID at connection time, and the requests will be handled according to that application profile.

        Every Bigtable Instance has a default application profile associated with it, this method configures the client to use it.

      • getAppProfileId

        public String getAppProfileId()
        Gets the app profile id that was previously set on this Builder.

      • setCredentialsProvider

        public BigtableDataSettings.Builder setCredentialsProvider​(com.google.api.gax.core.CredentialsProvider credentialsProvider)
        Sets the CredentialsProvider to use for getting the credentials to make calls with.

      • getCredentialsProvider

        public com.google.api.gax.core.CredentialsProvider getCredentialsProvider()
        Gets the CredentialsProvider to use for getting the credentials to make calls with.

      • setRefreshingChannel

        @Deprecated
public BigtableDataSettings.Builder setRefreshingChannel​(boolean isRefreshingChannel)
        Deprecated.
        Channel refreshing is enabled by default and this method will be deprecated.
        Configure periodic gRPC channel refreshes.

        This feature will gracefully refresh connections to the Cloud Bigtable service. This is a feature to address tail latency caused by the service dropping long lived gRPC connections, which causes the client to renegotiate the gRPC connection in the request path, which causes periodic spikes in latency.

      • isRefreshingChannel

        @Deprecated
public boolean isRefreshingChannel()
        Deprecated.
        Channel refreshing is enabled by default and this method will be deprecated.
        Gets if channels will gracefully refresh connections to Cloud Bigtable service.

      • getPrimingTableIds

        @Deprecated
public List<String> getPrimingTableIds()
        Deprecated.
        This field is ignored. If isRefreshingChannel() is enabled, warm up requests will be sent to all table ids of the instance.

      • enableBatchMutationLatencyBasedThrottling

        @BetaApi("Latency based throttling is not currently stable and may change in the future")
public BigtableDataSettings.Builder enableBatchMutationLatencyBasedThrottling​(long targetRpcLatencyMs)
        Enable latency based throttling for BigtableDataClient.newBulkMutationBatcher(String) with a target rpc latency. The number of allowed in-flight requests will be adjusted to reach the target bulk mutations rpc latency.

        The logic of adjusting in-flight request limits is as follows: 

         To start, Batcher allows FlowController.getCurrentElementCountLimit()
 in-flight elements with a total size of FlowController.getCurrentRequestBytesLimit().

 Every 20 seconds, Batcher checks the mean rpc latency of the requests and compare
 it with the target rpc latency:
   if (mean latency > 3 * target latency) {
     decrease element count limit by 30% of FlowController.getMaxElementCountLimit()
   } else if (mean latency > 1.2 * target latency) {
     decrease element count limit by 10% of FlowController.getMaxElementCountLimit()
   } else if (there was throttling in the past 5 minutes
        && mean latency < 0.8 * target latency) {
     increase element count limit by 5% of FlowController.getMaxElementCountLimit()
   } else if (there was throttling in the past 5 minutes
        && parallelism is 5% of FlowController.getMaxElementCountLimit()
        && mean latency < 2 * target latency) {
     increase element count limit by 2% of FlowController.getMaxElementCountLimit()

 Increases are capped by FlowController.getMaxElementCountLimit(), Decreases are floored at FlowController.getMinElementCountLimit() so that there is some level of throughput.
        See Also:
        for explanation on default configurations.

      • isLatencyBasedThrottlingForBatchMutationEnabled

        @BetaApi("Latency based throttling is not currently stable and may change in the future")
public boolean isLatencyBasedThrottlingForBatchMutationEnabled()
        Gets if latency based throttling is enabled for BigtableDataClient.newBulkMutationBatcher(String)

      • getTargetRpcLatencyMsForBatchMutation

        @BetaApi("Latency based throttling is not currently stable and may change in the future")
@Nullable
public Long getTargetRpcLatencyMsForBatchMutation()
        Gets target bulk mutation rpc latency if latency based throttling is enabled for BigtableDataClient.newBulkMutationBatcher(String). Otherwise returns null.

      • setBulkMutationFlowControl

        @InternalApi("Intended for use by the Bigtable dataflow connectors only")
public BigtableDataSettings.Builder setBulkMutationFlowControl​(boolean isEnableFlowControl)
        Configure flow control for BigtableDataClient.newBulkMutationBatcher(String) based on the current load on the Bigtable cluster.

        This is different from the FlowController that's always enabled on batch reads and batch writes, which limits the number of outstanding requests to the Bigtable server.

      • isBulkMutationFlowControlEnabled

        @InternalApi("Intended for use by the Bigtable dataflow connectors only")
public boolean isBulkMutationFlowControlEnabled()
        Gets if flow control is enabled for BigtableDataClient.newBulkMutationBatcher(String) based on the load of the Bigtable server.