*URGENT* clients can't connect to the WebCallServer - I get 502 from activation is it related ?

[Dani]

02:00:33,657 ERROR Subscription - Subscription Thread Ping failed
com.flashphoner.server.license.activation.A.A: Can not get response from activation server https://my.flashphoner.com/license/...&serverId=XXXXXXXXXXXXXXX&hostname=myhostname
at com.flashphoner.server.license.Activation.httpsRequest(Unknown Source)
at com.flashphoner.server.license.Activation.ping(Unknown Source)
at com.flashphoner.server.license.B.B.A(Unknown Source)
at com.flashphoner.server.license.B.B.C(Unknown Source)
at com.flashphoner.server.license.B.B.run(Unknown Source)

 

[camaro396]

+1 here. Restarting a service helped to keep it working but still I get these 502s in license.log.
Tried to cURL that URL and it worked OK.

 

[Max]

Good day.
We experienced an outage on our licensing servers. The issue has been resolved, and we’re investigating the root cause. Nowall the servers should work.

 

[Dani]

Thanks. I think a grace period should be applied so service will not die in such case - if license server cannot be reached let the server run for at least 72 hours before you "kill" the service, giving you enough time to recover without effecting running customers

 

[Max]

WCS sends a health check (a “ping”) to license servers endpoint every 15 minutes. If there is no response three times in a row, the license switches to Suspended. In practice, this provides a grace period of roughly 45–60 minutes during which the license continues to work without reaching the server.
The license server endpoint is a load balancer with a multiple backend servers behind.
We already implemented an automatic backend fallback if all the main backend servers go down. We also will add more servers behind the load balancer and/or add resources.
So the incidents like this should not repeat any more.