reactor-netty
reactor-netty copied to clipboard
reactor
Configurable request retry
Allow more configuration when it comes to retrying requests.
This is currently failing at least 1 test for 2 reasons that need to be addressed. First, Retry#max exhaustion generates the retryexhausted wrapper. I think that's a good thing, but I understand why you'd go a different route from multiple perspectives (IO checked retry, breaking behavior). Still thinking on that, but feel free to tell me what to do.
Also, in the error case, this fires the doOnError hook maxRetry + 1. Still looking into this.
Need to stop for the day, but figured I'd get a draft out there. I'll clean up all commits/code/etc. before moving out of draft.
#2425
@crankydillo Please sign the Contributor License Agreement!
Click here to manually synchronize the status of this Pull Request.
See the FAQ for frequently asked questions.
Thanks Violeta. Honestly, I need to work on it. It was not as easy as initially thought:) I can pass all the tests with my current code (not pushed yet), but I'm not happy about it, and I think there may be some corner cases I'm missing. I'll do my best to work on it a bit over the next week.
I pushed it, but I'm not happy about it. It's going to be a busy week or two, so I'm not sure when I will come back to it. Bottom line is I couldn't figure out how to avoid some huge redesign to the observers and still do this in a clean way. The crux of the problem revolves around getting that done-retrying signal to the observers at the right time. The only other idea I have is to have an internal, failing always sink-based flux that fails always embedded within HttpObserver, but that feels hacky, generates a bunch of unnecessary exceptions, and still has some problems. The main one being not knowing exactly what prompts retry (since its configurable!).
On a side note, I thought I'd already signed the Pivotal agreement. I'll get it signed again I guess.
@violetagg I'm just going to close this. I understand the desire to leverage reactor's retry, but I don't see how that can quickly be done. Given that my team is not demanding this feature, I doubt I'll be able to resolve this.