Right now, the GET operation that takes no ID path param, used by the Summary JSP in the UI, does an open-ended query of all accounts in the CouchDB database. If you had thousands, this would obviously be very expensive. Need to add a way to ask for the first 20 or whatever, with some explicit "order by", ensuring we get the same 20 that Portfolio would get from its PostgreSQL (or whatever relational database you picked), and then a way to advance to the next page of output, and so on. @rtclauss's performance tests in Istio/CloudWatch/Kiali suggest this "get all" to CouchDB is the current bottleneck in Looper stress runs.
Right now, the GET operation that takes no ID path param, used by the Summary JSP in the UI, does an open-ended query of all accounts in the CouchDB database. If you had thousands, this would obviously be very expensive. Need to add a way to ask for the first 20 or whatever, with some explicit "order by", ensuring we get the same 20 that Portfolio would get from its PostgreSQL (or whatever relational database you picked), and then a way to advance to the next page of output, and so on. @rtclauss's performance tests in Istio/CloudWatch/Kiali suggest this "get all" to CouchDB is the current bottleneck in Looper stress runs.