https://supertokens.com/ logo
Join Discord
Powered by
Why this error comes sometime randomly ? Not even ...
# support-questions-legacy
y
Why this error comes sometime randomly ? Not even letting me signIn again
r
hey @youscript
are you using our managed core service?
y
Yes
r
does this always happen?
y
No It just starts happening randomly
And randomly goes away too
r
hmm. im not sure that this is entirely related to supertokens.
are you making some custom request in the auth route?
y
Dont know but when this happens I cant seem to do any thing. Sending request to the api or signIn using super Tokens signIn func
r
i will need more info about this. For example, where exactly this is happening. The core stack above doesn't seem to indiciate it's from our SDK, therefore, you need to check.
y
Just to tell you that this error did't use to happen. Porcellus shared a branch with me some days before fixing the polyfill cross fetch error. The error has been coming since using that branch
r
Hmmm. Then maybe we will wait for @porcellus to help
y
Ya. I have seen the logs saying request timed out when refresging token. In the above logs, those logs are not there but I have seen those.
r
@porcellus will help when available 🙂
y
Yes, reassured it the error is only coming with this branch npm i supertokens-node@supertokens/supertokens-node#backwards/edge_compatibility_check and not with the prev version I was using which was 15.2.0
p
hi. so the error seems to be some kind of network issue: the "connection timout" in the logs basically mean that some kind of fetch call you made is timing out. It's not entirely sure, but it looks like you are not able to connect to the core, which is expected to produce errors.
r
If it’s not able to connect to the core, won’t the stack trace contain the querier file in the path? Or some path from supertokens lib?
p
it should, but it's not a guarantee. in older releases (before 16.0) async-await was transpiled away, which resulted in broken stacktraces.
this happens with other libraries as well.
r
I see
So this mean that the core is “taking time” to reply?
p
maybe. or some kind of connection issue
r
@youscript how often does this happen?
But it’s also strange that this is happening now and not before (before when the original branch was used)
p
so this may indeed be a fetch call we are making. It could be just a change in "reporting" and/or because the calls in the middleware is now actually working (and sometimes failing)
r
I highly doubt that it’s the core being slow here.
@youscript how often does this happen?
j
I've had an issue like that but it turned out I used a outsidethe button used to login for some reason it loads the UI before the backend is ever called and then fails to fetch. might help not sure
y
@rp_st It usually happens at the start. And then after some time gets fixed randomly. Then it usually stays ok for about an hour or 2 than starts happening again
and actually When I started using the branch when provided, the error did't use to come in the start and only used to come some time and gets fixed on a refresh or reload of the website. But today the error was coming in the start and was not even getting fixed on a refresh or reload.
@ja3nyc I dont think so this could be an issue because the error was not coming with the original branch. But will give it a try tho, maybe it will fix
r
Hmm. Maybe it’s worth setting up a debugging call for sometime next week for this? You could pick a time here: supertokens.com/call-user
y
https://github.com/yousuf123456/hanouts_makers Here's the repo incase you guys want to try reproducing the error. Just checked right now the error is not coming in the start now
r
Hmm. We won’t be able to check the repo unfortunately. But if you can reproduce it, then do setup a call so we can help - that will be the quickest way
y
Ya sure
p
Also, it could still be an issue not connected to supertokens - the fact that this only happens on the new branch doesn't mean it's connected supertokens: some functionality (e.g..: the middleware) wasn't really working, so getting past those errors you may be hitting new issues that may or may not be connected to our SDK
y
The issue did't come yesterday the whole day and not today until now. The error is coming now.
And pretty strangely goes away in 5 minutes. Now the error is again not coming
r
hmmm
maybe you could spin up a self hosted version of the core and see if that makes a diff? If it does, and if moving back to the cloud dev env, then we know that something is wrong there.
y
Will give it a try
Could you please give me highlevel overview of could I spin up a self hosted version of the core to make things quicker 😀?
r
PreviousNext
Powered by