Timeouts connecting to Azure API

providers
fine
gaprindashvili

#1

I’m running into an issue with MIQ timing out when refreshing Azure providers, seemingly all at once and on both Fine-4 and Gapri-3. Here is a gist with the traceback from Gapri-4. The behavior seems similar across the two versions though, with timeouts crashing the entire refresh process, and the timeouts not always coming from the same object, e.g. they will sometimes occur when polling deployment templates, sometimes when polling VM sizes.

Has anyone else run into this?


#2

This resolved itself spontaneously, and I believe was due to an issue on Azure’s end.


#3

Glad it worked itself out. It does bring up the issue of how to better handle timeouts in general, though. Maybe post that as an issue on the github tracker so we can flesh out the best way to deal with it.


#4

Can do. I was looking at the code while this was going on and I think you guys have an exception catcher that will attempt a retry on certain HTTP codes, hopefully it’ll be as simple as adding the timeout exception to that.