I ran into an issue and would like to get your input about how I can avoid this from happening.
Service A sends HTTP requests (
vertx-web-client version 4.2.4) to Service B (vertx-web-graphql version 4.2.4)Service B crashed
Service A started receiving timeouts the number of active connections spiked to 10x the amount of open connections
Service B recovered
Service A, managed to get some calls performed, but kept on failing, all requests that have failed did not end properly (custom router handlers were not called - Vert.x router logged -
Unhandled exception in router) The problem was resolved only by restarting Service A, even though it's vitals were ok. Service A did perform the requests using a retrier.
My question is - is there any configurations I could change in the WebClient to avoid this kind of issue? My understanding is that the failed connections were unstable or in some kind of unusable state - is there a way to reset them if it reaches this point?
Thanks