The compression works, the caching doesn't. It is possible that haproxy is messing up the etag which is invalidating the browser's cache?
On Tue, Jan 24, 2017 at 5:54 AM, Dennis Oelkers <[email protected]> wrote: > Hey Richard, > > thanks for the extensive and very valuable feedback. We put some thought > into architecting the web interface in a way that it's still usable on > slower connections/lower bandwidths and/or high-latency links, but it seems > we are still not there. Doing this is always a tradeoff between keeping > assets small, providing plugin authors a degree of freedom and providing a > rich interface. > > What I am wondering though, is that compression and caching does not work > for you. Both have been introduced for the web interface for a long time > now (see https://github.com/Graylog2/graylog2-server/commit/ > 515c3a242a3fdff1dcb3a4a2fece8351d0060c02 and https://github.com/ > Graylog2/graylog2-server/commit/68f0a0bc9508597cc5b81adc0c0b3496b5ac19eb). > Do you have any special configuration or components in your environment? > > Kind regards, > > D. > > On Wednesday, 4 January 2017 04:43:27 UTC+1, Richard S. Westmoreland wrote: >> >> Hello, >> >> I'm working on putting a GL cluster behind a reverse proxy and load >> balancer and while troubleshooting a url redirect issue I discovered why >> the interface has seemed really slow to me. My servers are in another >> country so latency becomes obvious for me. >> >> Just getting to the login screen are 9 requests totally 9.3 MB in >> transfer. That is huge. Individually it may be less of a problem but on >> the login screen the browser is preloading all these javascript files even >> when they aren't needed yet, and it is waiting until most of them are >> loaded before rendering the page. >> >> My short-term proposal: At UI login, don't preload. Just handle >> authentication first. Don't load certain JS files until the page that uses >> it requires it. >> >> My long-term proposal: I think the javascript is going to need some >> refactoring. For example, I'm looking at the MapWidget code and I see >> multiple sections of code concatenated together, which is easy to spot >> because of the sudden break of commented opensource license notices. If >> more of these are loaded individually, it may load faster since they will >> be separate/parallel requests. >> >> In the mapwidget code I see jQuery UI Widget 1.12.1, which is also being >> loaded by the enterprise widget, maybe more of them. So there is a lot of >> redundant code being downloaded and re-executed. >> >> Thanks for reading. >> >> Also is this the best place to post this? Or is there a dev/eng specific >> forum I can put these kinds of requests on? >> >> Thanks, >> Rick >> >> -- > You received this message because you are subscribed to the Google Groups > "Graylog Users" 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/graylog2/0d0cc565-f4b9-4b0a-8e77-cfc2c7f0b2f8%40googlegroups.com > <https://groups.google.com/d/msgid/graylog2/0d0cc565-f4b9-4b0a-8e77-cfc2c7f0b2f8%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 "Graylog Users" 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/graylog2/CAFnE7wPMyXQNCsnSoyYXPX6qN4UsY9YZLP9TvZQWLoshpHHoPg%40mail.gmail.com. For more options, visit https://groups.google.com/d/optout.
