presearchofficial / opensearch-frontier Goto Github PK

A Frontier instance can have many clients at the same time. In the current version of the service, the URLs are loaded from Opensearch whenever a client calls GetURLs and there aren't enough URLs cached for a given queue.

If many clients query the Frontier at the same time, this would result in a very large number of requests sent to Opensearch and a large latency for results to be returned to the clients.

Instead, we could have a separate thread for each service running in the background and querying opensearch for URLs to be returned to the clients. This way the rate of querying Opensearch would be controlled and constant. The client requests would also be served quicker as only URLs already obtained from Opensearch would be returned, if none or not enough are available, the service would return what it has in cache. This way the latency would be low, regardless of the number of clients connecting to the service.