Reed Allman

not sure i follow about moving w/o copying with the way the runnercall is done atm, if you want to have it at / explain, at least it sounds nicer...

no worries. the copy isn't too bad, I guess the sad part is we're copying in two places, but we can probably change that, too.

Hi @chryswoods, yea, this is expected. Agree, this would be a useful feature - of course, we expect most fn deployments to add some kind of auth on top of...

hmm. we have methods to add endpoints to the server itself and we have handlers (the invoke handler which uses the same var), so while this solves the problem I'm...

this is possible if fn is run with SSL certs (which turns on http/2) seamlessly. afaik, there is no reason to differentiate http versions at the trigger level and we...

we can document turning on h2 if it's not, it only works over https in go, but more complicated installs still wouldn't change the semantics here for triggers, so long...

this issue lacks a smoking gun, it could be anything... turn on debug logs and get a stack trace on program exit or something, please. it could be even your...

so, we talked about this a bit, but posting for posterity, but actual ram usage isn't indicative of whether your process is safe from the oom killer. it's a complex...

@denismakogon there should be a log message? there's also a metric, granted metrics is kind of an open ended issue in general. we're hoping to have all stuff like this...

Hi @irdara we don't have first-class support for OpenShift installation at this time, though it is almost certainly possible to accomplish somehow. We can leave this ticket open for future...