Incident Delay Attribution Experience API

Token Generation

Consumer endpoint needs to be used for generating token. The generated token will be valid for 3600 seconds and it needs to be passed as authorization token while initiating connection with Web Socket endpoint.

Caching of Messages

Messages are cached for 5 minutes to cover the temporary connection issues and will be send when consumer re-connects. Consumers have to pass a Sequence Id in a header (msgSequenceId) while re-connecting to the Web Socket. If Sequence Id is passed as 0 then consumer will receive all the messages cached for last 5 mins. If Sequence Id provided in the re-connection request was generated within last 5 mins then consumer will receive cached messages having Sequence Id greater than the Sequence Id provided in re-connection request. If the provided Sequence Id is 5 mins older then consumer will not get any cached messages and will only get the live feed.

Primary and Secondary Feed

There are two endpoints, where one endpoint acts as primary and the other as secondary. In case the primary endpoint is non-responsive, consumers have to switch over to the secondary endpoint.

Disconnections
Web Socket connection will timeout if idle for 60 minutes. Consumer has to implement the appropriate reconnection strategy to re-establish the connection in the event of any disconnection. There can be only one Web Socket connection open for given client application. If connection request received for the client application, which has existing Web Socket connection open then that existing connection will be closed before establishing the new connection.

Operational Characteristics

The API is designed to handle ~ 100k messages per day with message size < 100kb

Up to 10 messages per second at rush hour