[06:40:42] vgutierrez: awesome work with TLS 1.3! [06:46:06] <3 [08:17:32] 10Traffic, 10Operations, 10Goal, 10Performance-Team (Radar): Support TLSv1.3 - https://phabricator.wikimedia.org/T170567 (10Gilles) >>! In T170567#6062022, @Stashbot wrote: > {nav icon=file, name=Mentioned in SAL (#wikimedia-operations), href=https://sal.toolforge.org/log/g-SWgnEBv7KcG9M-vpmM} [2020-04-16T... [08:18:22] 10Traffic, 10Operations, 10Patch-For-Review: Memory leak on ats-tls 8.0.6 - https://phabricator.wikimedia.org/T249335 (10Gilles) >>! In T249335#6062023, @Stashbot wrote: > {nav icon=file, name=Mentioned in SAL (#wikimedia-operations), href=https://sal.toolforge.org/log/EfuXgnEBj_Bg1xd3h9Ty} [2020-04-16T10:44... [08:47:51] 10netops, 10Operations, 10homer, 10Patch-For-Review: Homer: Netbox driven switch interfaces - https://phabricator.wikimedia.org/T250429 (10ayounsi) [10:32:12] 10Traffic, 10MediaWiki-Cache, 10Operations, 10Page Content Service, and 3 others: cache_text cluster consistently backlogged on purge requests - https://phabricator.wikimedia.org/T249325 (10Urbanecm) >>! In T249325#6063043, @ema wrote: > @Urbanecm, @AntiCompositeNumber: esams has now been running with the... [11:38:17] 10Traffic, 10Operations, 10Goal, 10Performance-Team (Radar): Support TLSv1.3 - https://phabricator.wikimedia.org/T170567 (10Ladsgroup) That's weird, TLSv1.3 is famous for being faster than v1.2. [13:18:51] 10Traffic, 10Operations, 10Goal, 10Performance-Team (Radar): Support TLSv1.3 - https://phabricator.wikimedia.org/T170567 (10Gilles) In theory with features like 0-RTT resumption of course, but that doesn't mean that implementation, configuration and the real world follow suite with the theory. I don't know... [13:43:51] 10Traffic, 10Operations, 10Goal, 10Performance-Team (Radar): Support TLSv1.3 - https://phabricator.wikimedia.org/T170567 (10Vgutierrez) we haven't deployed 0-RTT at this time but even without it, a full TLSv1.3 handshake requires 1 RTT less than a full TLSv1.2 handshake. Thanks for the detailed report @Gil... [13:44:38] 10Traffic, 10Operations, 10Goal, 10Performance-Team (Radar): Support TLSv1.3 - https://phabricator.wikimedia.org/T170567 (10Ladsgroup) >>! In T170567#6065576, @Gilles wrote: > In theory with features like 0-RTT resumption of course, but that doesn't mean that implementation, configuration and the real worl... [13:53:48] nice work indeed valentin :) [13:54:16] well.. it looks like we need to dig some perf issues as gilles is pointing out [13:55:22] yes... but that doesn't take away from it being a big milestone :) [15:24:05] one question for atskafka - it seems that on Kafka I see a lot of things like [15:24:18] "http_method":"GET","http_status":"000" [15:24:38] with ip:"-" etc.. [15:24:44] and most of the time "accept":"image/webp,*/*" etc.. [15:25:02] are we doing some prefetch or similar in ATS? IIRC we had something similar for varnish [15:31:11] ah yes I think it was https://phabricator.wikimedia.org/T27611 [15:34:41] so in theory something like the follow happens [15:34:49] ats-tls gets a request for say a jpg [15:35:07] then varnish frontent gets it too, and tries webp [15:35:37] if it fails, it tries the original (stripping the .webp from the url) [15:36:07] but I am not sure why the retry bubbles up to ats-tls [15:36:17] or maybe something else is happening [17:25:13] 10Traffic, 10Analytics, 10Operations: Create replacement for Varnishkafka - https://phabricator.wikimedia.org/T237993 (10elukey) There is currently too much data that flows to kafka, for cp3050 we have 36GB * 12 partitions for a single day, definitely too much. I took a look to kafka messages and using `kaf... [18:04:06] 10netops, 10Operations, 10homer, 10Patch-For-Review: Homer: Netbox driven switch interfaces - https://phabricator.wikimedia.org/T250429 (10ayounsi)