maxRetransmits

maxRetransmits is an option used in the WebRTC RTCDataChannelInit configuration to control the reliability of messages sent over a data channel. It enables partial reliability by limiting the number of retransmissions that will be attempted for each message, trading delivery assurance for lower latency.

How it works: maxRetransmits is a non-negative integer N. For each message, the underlying transport will try to deliver the data up to N retransmissions. If the message cannot be delivered within those attempts, it is dropped and not delivered to the peer. A value of 0 means no retransmissions, i.e., best-effort delivery with no retry. maxRetransmits is mutually exclusive with maxPacketLifeTime; you should choose one mechanism to govern reliability, not both, as they express different retry policies.

Usage: To enable limited retransmission, the data channel can be created with a configuration such as { ordered: true, maxRetransmits: 5 }. This configuration allows up to five retransmission attempts per message. If time-based control is desired instead of a fixed retry count, maxPacketLifeTime can be used to specify a maximum time in milliseconds for retries, rather than a fixed count.

Compatibility and considerations: Partial reliability through maxRetransmits is supported by major WebRTC implementations, though exact behavior can vary by browser and version. This option is appropriate when low latency is prioritized over guaranteed delivery. For applications requiring strict delivery guarantees, avoid enabling maxRetransmits and use reliably delivered channels instead. See also RTCDataChannelInit, maxPacketLifeTime, and data channel reliability models.