Oliver De Guzman: Anyone know if we ever enabled facebook and paypal login for binglee? Or anyone know of a reason why they are disabled?

Lloyd Hazlett: they previously had google and my guess is they disabled it at some point recently after they’d reported it wasn’t working anymore: . Nigel probably forgot.

Oliver De Guzman: I see. He must have also forgotten that Facebook login was never enabled in the first place.

Oliver De Guzman: I don’t see any config changes disabling facebook login and there are no credentials configured for facebook login in the config

Oliver De Guzman: paypal login was disabled in 2016

Oliver De Guzman: they have an old version of fontis/sociallogin extension, even if I enabled facebook login on staging, I’m not able to login because facebook is not sending the email address as expected

Lloyd Hazlett: i’d say we just address google as per the original ticket

code review

I tried:
I think:

the closing brace and the elseif must be on the same line.

I’m not sure there’s any benefit to translating this text.

Peter Spiller: are there any logs we could import into elasticsearch? that way we could use kibana to browse them, look for patterns, etc.

Lloyd Hazlett: the logs are useless though, they are impossibly verbose

Lloyd Hazlett: we have to do something though, what we have just does not stack up

Lloyd Hazlett: it isn’t really production-grade

Peter Spiller: are there any rules we could apply to filter the existing logs?

Peter Spiller: or could we redirect what currently goes into emails to a log?

Lloyd Hazlett: we’ll have to wait for Ron and Oliver to update us

Ronilo Carr: those errors emails at the moment are the output from the systemd services which is basically a stack trace of a fatal error. we can’t really make those show a friendly error message can only turn down the log level which will make it even less useful.

Peter Spiller: are they useful for debugging?

Ronilo Carr: it’s only useful for letting us know when something has happened as we usually have to look up the entire log in the database

Ronilo Carr: if we turn off the errors emails completely from the systemd services we basically won’t know when jobs are constantly running into fatal errors

Ronilo Carr: all other errors are already being logged in cartwheel

Oliver De Guzman: For the pentaho order export, we can ignore the pentaho errors since we check for the output of the job — eg. we have an alert when recent orders are not being exported

Ronilo Carr: those errors from hype stg do need our intervention, looking into it