Authorization Extension fails with blocked event loop when signing-in

@jmangelo
Sorry for the delayed answer.
Our tenants are in the AU.
It seems we haven’t had any “Authorization Extension: Blocked event loop” exceptions logged in a while now (since the 31/10).
Should we consider this fixed?

We are in AU and last saw this on 31-Oct - the 2-Nov was the last time that have seen a ETIMEOUT error too - which we were getting frequently prior to that.

cheers,
dave

Thanks for the update, with regards to the blocked event loop yes, there are no outstanding actions and we are also treating the issue as resolved as we did not received any further reports about it.

The ETIMEOUT situation I believe was technically not the same situation although the other issue could likely increase that possibility so not seeing timeout anymore may have been a positive side-effect. However, there are certain cases where the extension may timeout because of load and as such based on what I know, although timeouts may still occur it’s not because of the original issue.

From jmangelo’s comment

Thanks for the update, with regards to the blocked event loop yes, there are no outstanding actions and we are also treating the issue as resolved as we did not received any further reports about it.

Do we have any update on this?

@saltuk We’re still experiencing this issue, any idea when a fix will be deployed?

An official update would be good here. We’re considering an implementation that would make use of the authorization extension, but seeing this gives us some pause to rely on the extension.

Do we have any update on this?

@saltuk We’re still experiencing this issue, any idea when a fix will be deployed?

An official update would be good here. We’re considering an implementation that would make use of the authorization extension, but seeing this gives us some pause to rely on the extension.

@grosendorf as far as I’m aware and I already mentioned in a comment there are no outstanding action on this; there was a fix applied and (already) deployed. The situation was monitored after the fix and there were no evidences of the issue so it was considered fix. From your comment it seems that you experience the issue recently so if you can include additional information that would be great. Frequency with which you get the error, if it’s specific to a single tenant, if you have any other extensions installed.

@odell see my comment below or on the question itself; the summary is that the situation was considered as fixed and until the mention by @grosendorf I was not aware of any other reports about this (after the fix). In relation to the extension, if you’re considering it, my advice would be for you to properly review it for your specific use case and also go through the reference documentation for some information about constraints.

We only have a single tenant, so yes, specific to just that one. We have 4 other extensions installed, plus a fair amount of custom rules. It’s difficult to give a solid answer as to the frequency, maybe every 20 user logins produces the error, definitely enough to be a serious issue for us.

I opened a ticket yesterday, and was told that the only workaround currently is to manually update the generated rule to increase the timeout. I was also told that “We are currently working on mitigating the slowness”, although that seems to conflict with @jmangelo 's comment above.

Technically it does not conflict, although you can argue with that because the visible outcome is similar; the situation is that the issue I was talking about was a systemic/infrastructure one that aggravated the overall situation and significantly increase the impact/frequency of this issue. That one was the one that I was treating as addressed. Having said that I just learned that the extension itself depending on usage may still lead to the same visible outcome and that’s why you were recommended to try to increase the timeout.

@saltuk - has this fix been rolled out to the EU Tenants - Im experiencing this issue daily, with usually between 5-6 occurrences. Attempting to reauthenticate straight away again works. Its getting quite frustrating.

The error reported by the Auth0js client is below:

Object {error: “unauthorized”, errorDescription: “Authorization Extension: Blocked event loop”, state: “0akMlR9tzcoczNOIiGdB1lwsqGZgMHcF”}
actions.js:18
error:“unauthorized”
errorDescription:“Authorization Extension: Blocked event loop”
state:“0akMlR9tzcoczNOIiGdB1lwsqGZgMHcF”
proto:Object {constructor: , defineGetter: , defineSetter: , …}
defineGetter:function defineGetter() { … }
defineSetter:function defineSetter() { … }
lookupGetter:function lookupGetter() { … }
lookupSetter:function lookupSetter() { … }
constructor:function Object() { … }
hasOwnProperty:function hasOwnProperty() { … }
isPrototypeOf:function isPrototypeOf() { … }
propertyIsEnumerable:function propertyIsEnumerable() { … }
toLocaleString:function toLocaleString() { … }
toString:function toString() { … }
valueOf:function valueOf() { … }
proto:null

Just confirming that we are NOT seeing this issue from Australia.

Hi again,

We have identified the reason for the remaining issue in EU tenants and applied a patch. Since 24 hours, we haven’t observed the same error across all EU tenants.

We will keep monitoring the clusters, but I can safely say that issue is now also resolved for EU tenants.

Apologies again for the inconvenience this has caused .

Hi again,

We have identified the reason for the remaining issue in EU region and applied a patch. Since 24 hours, we haven’t observed the same error across all EU tenants.

We will keep monitoring the clusters, but I can safely say that issue is now resolved for all regions including EU.

Apologies again for the inconvenience this has caused .

"Authorization Extension: Blocked event loop" error · Issue #165 · auth0/auth0-authorization-extension · GitHub – but no response there either.