699.0.0

@metamask/network-controller 27.0.0

Added

• Add NetworkController:rpcEndpointChainAvailable messenger event (#7166)
  • This is a counterpart to the (new) NetworkController:rpcEndpointChainUnavailable and NetworkController:rpcEndpointChainDegraded events, but is published when a successful request to an endpoint within a chain of endpoints is made either initially or following a previously established degraded or unavailable status.

Changed

• BREAKING: Split up and update payload data for NetworkController:rpcEndpointDegraded and NetworkController:rpcEndpointUnavailable (#7166)
  • NetworkController:rpcEndpointDegraded and NetworkController:rpcEndpointUnavailable still exist and retain the same behavior as before.
  • New events are NetworkController:rpcEndpointChainDegraded and NetworkController:rpcEndpointChainUnavailable, and are designed to represent an entire chain of endpoints. They are also guaranteed to not be published multiple times in a row. In particular, NetworkController:rpcEndpointChainUnavailable is published only after trying all of the endpoints for a chain and when the underlying circuit for the last endpoint breaks, not as each primary's or failover's circuit breaks.
  • The event payloads have been changed:
    • For individual endpoint events (NetworkController:rpcEndpointUnavailable, NetworkController:rpcEndpointDegraded): failoverEndpointUrl has been removed, and primaryEndpointUrl has been added. In addition, networkClientId has been added to the payload.
    • For chain-level events (NetworkController:rpcEndpointChainUnavailable, NetworkController:rpcEndpointChainDegraded, NetworkController:rpcEndpointChainAvailable): These include chainId, networkClientId, and event-specific fields (e.g., error, endpointUrl) but do not include primaryEndpointUrl. Consumers can derive endpoint information from the networkClientId using NetworkController:getNetworkClientById or NetworkController:getNetworkConfigurationByNetworkClientId.
• BREAKING: Rename and update payload data for NetworkController:rpcEndpointRequestRetried (#7166)
  • This event is now called NetworkController:rpcEndpointRetried.
  • The event payload has been changed as well: failoverEndpointUrl has been removed, and primaryEndpointUrl has been added. In addition, networkClientId and attempt have been added to the payload.
• BREAKING: Update AbstractRpcService/RpcServiceRequestable to remove { isolated: true } from the onBreak event data type (#7166)
  • This represented the error produced when isolate is called on a Cockatiel circuit breaker policy. This never happens for our service (we use isolate internally, but this error is suppressed and cannot trigger onBreak)
• Move peer dependencies for controller and service packages to direct dependencies (#7209)
  • The dependencies moved are:
    • @metamask/error-reporting-service (^3.0.0)
  • In clients, it is now possible for multiple versions of these packages to exist in the dependency tree.
    • For example, this scenario would be valid: a client relies on @metamask/controller-a 1.0.0 and @metamask/controller-b 1.0.0, and @metamask/controller-b depends on @metamask/controller-a 1.1.0.
  • Note, however, that the versions specified in the client's package.json always "win", and you are expected to keep them up to date so as not to break controller and service intercommunication.
• Automatically update network status metadata when chain-level RPC events are published (#7186)
  • NetworkController now automatically subscribes to NetworkController:rpcEndpointChainUnavailable, NetworkController:rpcEndpointChainDegraded, and NetworkController:rpcEndpointChainAvailable events and updates the corresponding network's status metadata in state when these events are published.
  • This enables real-time network status updates without requiring explicit lookupNetwork calls, providing more accurate and timely network availability information.