quic: allow port migration to config the number of timeouts needed to trigger path degrading (#18801)

Signed-off-by: Renjie Tang <renjietang@google.com>

Mirrored from https://github.com/envoyproxy/envoy @ 41d7be65863ef4f83cbd35eded2ff2fa798006ca
pull/626/head
data-plane-api(Azure Pipelines) 3 years ago
parent 6b0017c714
commit 3ca8331a51
  1. 7
      envoy/config/core/v3/protocol.proto

@ -83,15 +83,12 @@ message QuicProtocolOptions {
google.protobuf.UInt32Value initial_connection_window_size = 3
[(validate.rules).uint32 = {lte: 25165824 gte: 1}];
// [#not-implemented-hide:] Hiding until timeout config is supported.
// The number of timeouts that can occur before port migration is triggered for QUIC clients.
// This defaults to 1. If sets to 0, port migration will not occur.
// This defaults to 1. If set to 0, port migration will not occur on path degrading.
// Timeout here refers to QUIC internal path degrading timeout mechanism, such as PTO.
// This has no effect on server sessions.
// Currently the value can only be 0 or 1.
// TODO(renjietang): Plumb through quiche to make this config able to adjust the amount of timeouts needed to trigger port migration.
google.protobuf.UInt32Value num_timeouts_to_trigger_port_migration = 4
[(validate.rules).uint32 = {lte: 1 gte: 0}];
[(validate.rules).uint32 = {lte: 5 gte: 0}];
// Probes the peer at the configured interval to solicit traffic, i.e. ACK or PATH_RESPONSE, from the peer to push back connection idle timeout.
// If absent, use the default keepalive behavior of which a client connection sends PINGs every 15s, and a server connection doesn't do anything.

Loading…
Cancel
Save