Windy Authorization Offline
Even though this has been resolved, it would be useful to acknowledge that the AUTH was unavailable this morning. It might help others with any support cases that came in.:
This morning between around 11am GMT and 11:37 the Windy API Auth went offline.
In Chrome browsers this was visible in the console as:
Access to XMLHttpRequest at 'https://api.windy.com/api/map-forecast/v2/auth' from origin X has been blocked by CORS policy: No 'Access-Control-Allow-Origin' header is present on the requested resource.
In Fiddler, it was shown as:
The server encountered a temporary error and could not complete your request.Please try again in 30 seconds.
At 11:37 GMT the Auth began working again.
@CompassDSNM Hello, the reported issue is not active anymore. But thank you for the report.
How are we able to catch this error?
@dirtgambit This error should not occur anymore.
With the Windy Auth API offline it caused catastrophic issues to our application, basically making it totally unusable. We need to be able to handle this so if it happens again our web app can continue to work, but without Windy.
Could your developers please let us know how best to identify if the Auth API is not working (the same as yesterday) so that we can put in some defensive code?
@CompassDSNM Hello, there is no way how to identify this. We have updated our systems, causing an outage.
Hi Korina, we need to be able to defend the rest our site when it does happen again. There must be some way we can identify it and do something else?
@CompassDSNM Hello, the issue does not longer occur, there is no need for identification.
Hi Korina, we are still waiting for a proper response.
We know your service is currently working, but if it does ever go down again we don't want our whole site to be taken down with it.
So we kindly are asking you how we can detect Windy being down in the future so we can put in defensive code to handle it.
@CompassDSNM Hi, as @Korina said, that was an extreme situation and our services normally work as usual. Of course, it can happen at some point again. We do not guarantee 100% access to our services. There is no health check endpoint, so I am affraid you have to handle this on your own unfortunately. Sorry about that. Thanks to your recomendation, we will think about some health check in the future.