rpc: Add experimental config params to allow for subscription buffer size control (tm v0.34.x) (#7230)

A workaround for #6729. Add parameters to control buffer sizes for
event subscription RPC clients. On some networks, buffering causes
clients to be dropped and/or events to be lost.

For additional context, see the discussion on #7188.

- Add experimental_subscription_buffer_size config parameter
- Add experimental_websocket_write_buffer_size config parameter
- Add experimental_close_on_slow_client config parameter

Co-authored-by: M. J. Fromberger <fromberger@interchain.io>

config/config.go

@@ -52,6 +52,9 @@ var (
 
 	defaultNodeKeyPath  = filepath.Join(defaultConfigDir, defaultNodeKeyName)
 	defaultAddrBookPath = filepath.Join(defaultConfigDir, defaultAddrBookName)
+
+	minSubscriptionBufferSize     = 100
+	defaultSubscriptionBufferSize = 200
 )
 
 // Config defines the top level configuration for a Tendermint node
@@ -342,6 +345,29 @@ type RPCConfig struct {
 	// to the estimated maximum number of broadcast_tx_commit calls per block.
 	MaxSubscriptionsPerClient int `mapstructure:"max_subscriptions_per_client"`
 
+	// The number of events that can be buffered per subscription before
+	// returning `ErrOutOfCapacity`.
+	SubscriptionBufferSize int `mapstructure:"experimental_subscription_buffer_size"`
+
+	// The maximum number of responses that can be buffered per WebSocket
+	// client. If clients cannot read from the WebSocket endpoint fast enough,
+	// they will be disconnected, so increasing this parameter may reduce the
+	// chances of them being disconnected (but will cause the node to use more
+	// memory).
+	//
+	// Must be at least the same as `SubscriptionBufferSize`, otherwise
+	// connections may be dropped unnecessarily.
+	WebSocketWriteBufferSize int `mapstructure:"experimental_websocket_write_buffer_size"`
+
+	// If a WebSocket client cannot read fast enough, at present we may
+	// silently drop events instead of generating an error or disconnecting the
+	// client.
+	//
+	// Enabling this parameter will cause the WebSocket connection to be closed
+	// instead if it cannot read fast enough, allowing for greater
+	// predictability in subscription behaviour.
+	CloseOnSlowClient bool `mapstructure:"experimental_close_on_slow_client"`
+
 	// How long to wait for a tx to be committed during /broadcast_tx_commit
 	// WARNING: Using a value larger than 10s will result in increasing the
 	// global HTTP write timeout, which applies to all connections and endpoints.
@@ -391,7 +417,9 @@ func DefaultRPCConfig() *RPCConfig {
 
 		MaxSubscriptionClients:    100,
 		MaxSubscriptionsPerClient: 5,
+		SubscriptionBufferSize:    defaultSubscriptionBufferSize,
 		TimeoutBroadcastTxCommit:  10 * time.Second,
+		WebSocketWriteBufferSize:  defaultSubscriptionBufferSize,
 
 		MaxBodyBytes:   int64(1000000), // 1MB
 		MaxHeaderBytes: 1 << 20,        // same as the net/http default
@@ -425,6 +453,18 @@ func (cfg *RPCConfig) ValidateBasic() error {
 	if cfg.MaxSubscriptionsPerClient < 0 {
 		return errors.New("max_subscriptions_per_client can't be negative")
 	}
+	if cfg.SubscriptionBufferSize < minSubscriptionBufferSize {
+		return fmt.Errorf(
+			"experimental_subscription_buffer_size must be >= %d",
+			minSubscriptionBufferSize,
+		)
+	}
+	if cfg.WebSocketWriteBufferSize < cfg.SubscriptionBufferSize {
+		return fmt.Errorf(
+			"experimental_websocket_write_buffer_size must be >= experimental_subscription_buffer_size (%d)",
+			cfg.SubscriptionBufferSize,
+		)
+	}
 	if cfg.TimeoutBroadcastTxCommit < 0 {
 		return errors.New("timeout_broadcast_tx_commit can't be negative")
 	}

config/toml.go

@@ -206,6 +206,33 @@ max_subscription_clients = {{ .RPC.MaxSubscriptionClients }}
 # the estimated # maximum number of broadcast_tx_commit calls per block.
 max_subscriptions_per_client = {{ .RPC.MaxSubscriptionsPerClient }}
 
+# Experimental parameter to specify the maximum number of events a node will
+# buffer, per subscription, before returning an error and closing the
+# subscription. Must be set to at least 100, but higher values will accommodate
+# higher event throughput rates (and will use more memory).
+experimental_subscription_buffer_size = {{ .RPC.SubscriptionBufferSize }}
+
+# Experimental parameter to specify the maximum number of RPC responses that
+# can be buffered per WebSocket client. If clients cannot read from the
+# WebSocket endpoint fast enough, they will be disconnected, so increasing this
+# parameter may reduce the chances of them being disconnected (but will cause
+# the node to use more memory).
+#
+# Must be at least the same as "experimental_subscription_buffer_size",
+# otherwise connections could be dropped unnecessarily. This value should
+# ideally be somewhat higher than "experimental_subscription_buffer_size" to
+# accommodate non-subscription-related RPC responses.
+experimental_websocket_write_buffer_size = {{ .RPC.WebSocketWriteBufferSize }}
+
+# If a WebSocket client cannot read fast enough, at present we may
+# silently drop events instead of generating an error or disconnecting the
+# client.
+#
+# Enabling this experimental parameter will cause the WebSocket connection to
+# be closed instead if it cannot read fast enough, allowing for greater
+# predictability in subscription behaviour.
+experimental_close_on_slow_client = {{ .RPC.CloseOnSlowClient }}
+
 # How long to wait for a tx to be committed during /broadcast_tx_commit.
 # WARNING: Using a value larger than 10s will result in increasing the
 # global HTTP write timeout, which applies to all connections and endpoints.