@Vlad_GitHub "Errors in those services triggered a client-side retry loop that increased traffic during recovery. We had to mitigate that behavior before we could safely restore traffic" - so no circuit breakers yet ?
@system_monarch adaptive algorithms running on client are used to notify swiggy servers through WS , e.g. faster the speed of driver more frequent is the update , vice-versa and some other variables as well in algorithm to determine whether its worth to notify server for location update
@system_monarch No , ideally notification should be in your queue (kafka/sqs) and request to firebase shoumd be throttled by the rate of consumption of your worker from queue
@anshika_098 go for a walk -> reattempt -> if still cant solve -> then time to look into hints and asses where the actual gap was -> fill the gap -> AC next time if similar problem comes
@kav_kavi11 -> Kafka (different topics for different notification+prriority) -> add persistence for client to check status -> woker -> 3rd party provider [Firebase/APN/SendGrid etc]