As we already store the results in redis for each domain, we can just return that result early in the api pipeline. If we do, we can remove the $needsReply from rabbitmq and put that in as a low priority request. Hopefully this saves a lot of rabbitmq channels.
Another optimisation later on might be to put/check the time the last request was checked and only do an update if more than x hours ago. If we still care about the logging we could even fire a log request without a check request and just log the same result as we had from cache.
As we already store the results in redis for each domain, we can just return that result early in the api pipeline. If we do, we can remove the $needsReply from rabbitmq and put that in as a low priority request. Hopefully this saves a lot of rabbitmq channels.
Another optimisation later on might be to put/check the time the last request was checked and only do an update if more than x hours ago. If we still care about the logging we could even fire a log request without a check request and just log the same result as we had from cache.
lets first try this.