poorejc commented on issue #50: URL: https://github.com/apache/incubator-flagon-useralejs/issues/50#issuecomment-808621268
@confusingstraw you've always been someone I consider imminently sane. Sounds like the right way to go is [3] Websockets @confusingstraw makes a compelling case. - it solves the connectivity and event handling problems - it's something that's pretty ubiquitous - it aligns with previous user interest - the consequences of the "false positive" is that everything in the log queue is sent to back-end and queue is flushed. That results in neither duplicate logs or dropped logs. Seems like a low severity risk - the big drawback is that we have to manage/maintain supporting plugins for ELK and probably provide some other examples for plugins to a few other datastores (e.g., postgress). That's not too crazy. [1,2,4,5] seem like reaches and [5] seems like it would introduce other headaches. If nothing else heard (feel free to add opinion here @kevbrowndev, @UncleGedd), I can put together some new tickets to cover this. Could use some help on this one. Other thing would like opinions on: To do this right, we should also make sure to mod and test our ELK stack and run full end-to-end integration tests. That could take a little longer. Is it worth doing a little more testing on this branch as is, pushing into next minor release (2.2.0) as a 'hot-fix', then doing the refactoring on sendLogs to websockets as the following release (2.3.0). Would be nice to make sure than NPM users have something incrementally better than current sendOnClose behavior (which doesn't behave at all?) Opinions? -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org
