We should make sure that the Event Stream does not negatively impact Kibana performance. The main concern are the read events, as other CRUD events have much smaller volume.
We could either performa benchmarking for our read event ingestion.
Or, remove the problem altogether:
Do not track read events, at least, in the initial release.
Higher buffer for those high-frequence events. For example, buffer all read events for a minute, and send a _bulk write request only once a minute.
We should make sure that the Event Stream does not negatively impact Kibana performance. The main concern are the
read
events, as other CRUD events have much smaller volume.read
event ingestion.read
events, at least, in the initial release.read
events for a minute, and send a_bulk
write request only once a minute.