I uploaded a sample of buzzes from near the peak of yesterday's spike period to this group. I didn't dig too far in, but you'll prolly have a better idea of what to look for. nothing jumped right out at me.
Jud On Mar 4, 12:10 pm, Brett Slatkin <[email protected]> wrote: > Hey Jud, > > On Thu, Mar 4, 2010 at 10:36 AM, Jud Valeski <[email protected]> wrote: > > We see it every day... clockwork. > > Yeah it's interesting. I wonder if this is the combination of Western > Europe and the east coast waking up, combined with the crawling > schedule of Google Buzz (ie, resyndication of feeds from slower > providers like Twitter). Do you have any visibility into the > type/source of the Buzz updates that are present in that spike? I'll > have to dig a bit to find that. > > -Brett > > > On Wed, Mar 3, 2010 at 5:28 PM, Brett Slatkin <[email protected]> wrote: > > >> Hey Jud, > > >> On Wed, Mar 3, 2010 at 8:20 AM, Jud <[email protected]> wrote: > >> > We're seeing something strange with inbound POSTs from google hub > >> > subscribed buzz publisher > >> > user post topics. We have a "large" number of topics we're subscribed > >> > to, and everyday around 4am Eastern we see an approx. 3x spike in > >> > inbound POST volume that lasts for a couple of hours. > > >> > Our edge HTTP server logs are logging the flux, so we're confident > >> > (but not 100%) that our app isn't backing up. > > >> > As anyone who's played with queues before is aware... they can be a > >> > bear to chase down. I thought I'd post here just to see if anyone else > >> > is seeing similar behavior. > > >> > I posted this on the google buzz group with a slightly different > >> > flavor but haven't heard anything on that front. > > >> Thanks for the report. > > >> I do see a delivery spike on the reference hub from 1-2am PST last > >> night too. Not sure what the cause is. Please let me know if you see > >> this again. I'm trying to track down if this was overactive retries of > >> delivery or something else. > > >> -Brett
