Thanks Traun! That is very interesting data. I've got a few follow up questions:
- Will the number of pushers and pullers directly impact the lag time? - If so, could we achieve lag times in the neighborhood of 2 seconds if the number of pushers and pullers are kept to a limit? - Would I be able to run these tests myself and play around with the parameters? If so, any pointers to how to get it up and running would be a huge help. Thanks again! -- Eric On Monday, February 9, 2015 at 3:24:39 PM UTC-5, Traun Leyden wrote: > > Hey Eric, > > Sorry for the late reply. > > If you look at this performance run result > <http://ci.sc.couchbase.com/view/sync%20gateway/job/cen006-p0-cbsgw-real-perf/644/parameters/+we+are>, > > which is being run against a recent snapshot of the master branch, it's > running with 5K "pushers" and 8K "pulllers" (simulated clients pushing docs > and doing continuous pull replications), and we're seeing lag times of > approx 16.9 seconds -- meaning the time the doc was pushed to Sync Gateway > by a pusher, versus the time it appeared on the changes feed of a puller. > > That particular data point can be seen by looking at the > gateload_expvar2.json > <http://ci.sc.couchbase.com/view/sync%20gateway/job/cen006-p0-cbsgw-real-perf/644/artifact/gateload_expvar_2.json> > > file, and looking at this field: > > > > - PushToSubscriberInteractive: > { > - *...* > - p99: 16911357990 > }, > > > > which is in nanoseconds, so converting to seconds == 16.9 seconds. > > HTH > > > On Tue, Feb 3, 2015 at 6:05 AM, Eric Levine <[email protected] > <javascript:>> wrote: > >> Traun - What I'm really interested in trying to do is get an idea of the >> Sync Gateway's performance when using Couchbase as the database. Can you >> share some of your results from the perf tests you run? What I'd like to >> know is the time it takes to sync a document between clients, and from the >> REST API down to a client. Obviously, there area a lot of variables >> involved, but it would help to have something to set expectations. >> >> >> On Monday, February 2, 2015 at 6:58:07 PM UTC-5, Traun Leyden wrote: >> >>> >>> >>> On Thu, Jan 29, 2015 at 9:33 PM, Jens Alfke <[email protected]> wrote: >>> >>>> >>>> I'm not knowledgeable on deploying cloud-hosted Sync Gateways, so I >>>> don't have any idea how to troubleshoot this. Traun's done the most work >>>> on >>>> that, I think, so hopefully he can chime in. >>>> >>>> >>> I don't have much to chime in with, because I haven't done any perf >>> testing yet on the cloud-hosted Sync Gateways. >>> >>> For our internal "perf runner" tests we have at Couchbase, we don't >>> currently benchmark against Walrus, but I will suggest that we add that to >>> our test matrix. >>> >>> >>> -- >> You received this message because you are subscribed to the Google Groups >> "Couchbase Mobile" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to [email protected] <javascript:>. >> To view this discussion on the web visit >> https://groups.google.com/d/msgid/mobile-couchbase/39adaa99-ad87-45ad-ab12-27e5181f93e6%40googlegroups.com >> >> <https://groups.google.com/d/msgid/mobile-couchbase/39adaa99-ad87-45ad-ab12-27e5181f93e6%40googlegroups.com?utm_medium=email&utm_source=footer> >> . >> >> For more options, visit https://groups.google.com/d/optout. >> > > -- You received this message because you are subscribed to the Google Groups "Couchbase Mobile" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/mobile-couchbase/fe05366e-4748-4350-937c-ccc6dc458711%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.
