It would be great if someone from Google could clarify if today's behavior is what we can expect out of task queues and cron. I noticed that depending on when you refresh the App Engine status page today http://code.google.com/status/appengine you do see "Elevated" for task queues (and at other times, Normal).
If this flavor of Elevated "may happen a few times a week, deal", I'll need to figure out a workaround or look at hosting at least the non- user-facing part of my app elsewhere. Maybe resident backend instances scanning the DB for jobs to run...ughh..seems like that would get expensive would be more cumbersome to tune than task queues. On Jun 6, 4:51 pm, tempy <[email protected]> wrote: > Tom - I don't really see why heavy GAE-wide load should impact task > queues or cron jobs in this way. The tasks run within our existing > instances, and each instance, once it exists, has reserved a slice of > resources for its own use. If that instance is lightly loaded (as > mine are), it should be able to handle the task just fine, just as it > would a regular request - isn't that the whole point of "web hooks"? > The tasks are no different from ordinary requests, as far as the > instances are concerned. I always assumed that the way system-wide > load would manifest would be in the unavailability of new instances > for scaling. Of course, these are educated guesses and I'm probably > wrong. > > And I share your worries - my app is heavily asynchronous and large > chunks of it come to a screeching halt when the tasks stop operating - > I know task queues are advertised as "time of execution is not > guaranteed," but I hope that this does not include "tasks should > execute immediately, but may execute in ten minutes, or an hour, or > tomorrow." If that's the case I'm really on the wrong platform. > > On Jun 6, 10:38 pm, Tom Phillips <[email protected]> wrote: > > > > > > > > > Yes, same problem here. This happened during the day a couple days > > last week too. Tasks sit waiting to be dispatched even though none are > > running and the frontend instances are lightly loaded with low > > latency. A task runs once in a while, sporadically. > > > Cron jobs (once per minute in my case) also fail, only succeeding one > > in a while. They are "on time Failed", and my handler isn't even > > invoked when they fail. I'm assuming cron is using the same task > > infrastructure. > > > Raised production > > issuehttp://code.google.com/p/googleappengine/issues/detail?id=5150 > > on this. > > > App engine does a great job scaling user-facing requests, but I'm > > starting to wonder if when App Engine as a system overall is under > > heavy load it is doing so at the expense of non-user-facing requests. > > My app is lightly loaded now, but that will change - do I need to be > > worried that if my app is heavily asynchronous, with throughput and > > timeliness depending on smooth task queue operation, I will run into > > big problems at scale on days like today? > > > On Jun 6, 1:58 pm, tempy <[email protected]> wrote: > > > > Tasks for my app (ID: openkleio) are not executing, just sitting there > > > and executing every couple of minutes, sometimes. I can run them > > > manually just fine from the control panel. No quota issues, no errors > > > - just sitting there. Anyone else seeing this? -- 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.
