ResilientWebSocketClientOptions

All configuration for ResilientWebSocketClient lives on Options, a ResilientWebSocketClientOptions instance.

Property	Type	Default
BackoffStrategy	BackoffStrategy (None, Linear, Exponential)	None
InitialReconnectDelay	TimeSpan	1 second
MaxReconnectDelay	TimeSpan	30 seconds
SendQueueCapacity	int	256

Setting options

Either construct the options object directly and pass it to AsResilient:

var options = new ResilientWebSocketClientOptions
{
    BackoffStrategy = BackoffStrategy.Exponential,
    InitialReconnectDelay = TimeSpan.FromSeconds(2),
    MaxReconnectDelay = TimeSpan.FromMinutes(1),
};

var resilient = await client.AsResilient(uri, options).StartAsync();

or configure inline via the other AsResilient overload:

var resilient = await client
    .AsResilient(uri, o => o.BackoffStrategy = BackoffStrategy.Linear)
    .StartAsync();

BackoffStrategy, InitialReconnectDelay, and MaxReconnectDelay can also be set after construction via fluent extensions, see Fluent Extensions. SendQueueCapacity cannot, see below.

BackoffStrategy

How long to wait between reconnect attempts after a dropped connection.

None (the default)

No delay between reconnect attempts, retries immediately. This is deliberately the default: it's a real behavior, and a real risk (hammering a server that's genuinely down), rather than a hidden fixed delay, so using this in production requires actively choosing Linear or Exponential rather than inheriting an assumption made on your behalf.

Linear

Delay grows linearly with each attempt: InitialReconnectDelay * attempt (1s, 2s, 3s, 4s... with the defaults), capped at MaxReconnectDelay.

Exponential

Delay doubles with each attempt: InitialReconnectDelay * 2^(attempt - 1) (1s, 2s, 4s, 8s... with the defaults), capped at MaxReconnectDelay.

The actual calculation is exposed as a pure, static function independent of any live connection, ReconnectBackoff.GetDelay(BackoffStrategy strategy, TimeSpan initialDelay, TimeSpan maxDelay, int attempt), if you need to reason about or test the delay sequence directly.

SendQueueCapacity

The maximum number of not-yet-sent messages buffered while disconnected. Unlike the other three properties above, this one is only ever read once, at construction, to size the internal send queue. Changing it afterward, even immediately after construction, has no effect, which is exactly why there's no WithSendQueueCapacity fluent extension. Set it via the AsResilient(uri, Action<ResilientWebSocketClientOptions>) overload shown above, which configures the options before the client is actually built.