Update: After removing all of the memcache optimizations, the application is back up and running. So, it does appear that my performance problems are exclusively memcache related.
Is it reasonable to expect a memcache.get or a command like this: event = memcache.get_multi( ["time","cmd","data","targetid","playerid","roomid"] , key_prefix=key) To take between .5 and .9 seconds regularly? Even if it's accessing a key that's invalid, not until Oct 31st have a seen performance problems like this. If someone from Google or anyone else would like to see this issue in a live environment, this version is still using memcache: http://363.latest.gas.appspot.com/play/ I'm very curious to hear if anyone else is experiencing similar problems. Thanks. Rob On Oct 31, 11:12 pm, Rob <[EMAIL PROTECTED]> wrote: > I'm trying to narrow down some recent problems with our app > "gas" (http://www.guessasketch.com). For the last month or so, it has > been very stable. It's smoothly handled 10 requests per second during > busy times (which have been lasting sometimes for hours at a time). > Just in the last 15 hours the app is performing extremely poorly, and > not responding even under light load. In trying to debug, I'm > noticing particularly long memcache delays. > > For example, here is a snippet of code that appears to have terrible > performance tonight (in this case, currentEventId is none): > > logging.debug("Getting events for user" ) > if currentEventId is None: > currentEventId = memcache.get("Event:currentEventId") > logging.debug("-----User %s is starting with event #%s-------" > % (self.name(), currentEventId)) > > And here is the associated debug entries from tonight's log where it > appears that it's taking almost a full second to do a single > memcache.get between two entries to the log: > > #D 10-31 10:56PM 47.547 > Getting events for user > #D 10-31 10:56PM 48.479 > -----User is test starting with event #227------- > > Also, I've noticed the memcache seeming to lose variables (like this > integer) several times an hour which is much more frequent than I've > experienced in the past month. I'm not sure if it's related, but I'm > also getting a very high percentage of application reloads where the > python main function has to start from scratch in the log. > > I'd be very interested to hear if others are seeing similar problems > or if you have any other debugging suggestions for me to try to narrow > this issue down. > > Thanks in advance for any help. > > -Rob --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "Google App Engine" group. To post to this group, send email to [email protected] To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/google-appengine?hl=en -~----------~----~----~----~------~----~------~--~---
