From owner-freebsd-hackers@freebsd.org Tue Dec 13 16:31:36 2016 Return-Path: Delivered-To: freebsd-hackers@mailman.ysv.freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:1900:2254:206a::19:1]) by mailman.ysv.freebsd.org (Postfix) with ESMTP id 04C41C76939 for ; Tue, 13 Dec 2016 16:31:36 +0000 (UTC) (envelope-from slw@zxy.spb.ru) Received: from zxy.spb.ru (zxy.spb.ru [195.70.199.98]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (Client did not present a certificate) by mx1.freebsd.org (Postfix) with ESMTPS id B9DB7192D; Tue, 13 Dec 2016 16:31:35 +0000 (UTC) (envelope-from slw@zxy.spb.ru) Received: from slw by zxy.spb.ru with local (Exim 4.86 (FreeBSD)) (envelope-from ) id 1cGpzC-0001HK-MT; Tue, 13 Dec 2016 19:31:26 +0300 Date: Tue, 13 Dec 2016 19:31:26 +0300 From: Slawa Olhovchenkov To: Mark Rushakoff Cc: Ed Schouten , FreeBSD Hackers , Allan Jude , Jonathan Anderson Subject: Re: Sysctl as a Service, or: making sysctl(3) more friendly for monitoring systems Message-ID: <20161213163126.GA98176@zxy.spb.ru> References: <200A5D70-24CE-4843-B99D-E10A2AF4F68F@FreeBSD.org> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.5.24 (2015-08-30) X-SA-Exim-Connect-IP: X-SA-Exim-Mail-From: slw@zxy.spb.ru X-SA-Exim-Scanned: No (on zxy.spb.ru); SAEximRunCond expanded to false X-BeenThere: freebsd-hackers@freebsd.org X-Mailman-Version: 2.1.23 Precedence: list List-Id: Technical Discussions relating to FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 13 Dec 2016 16:31:36 -0000 On Tue, Dec 13, 2016 at 08:16:57AM -0800, Mark Rushakoff wrote: > In Prometheus' node_exporter, it looks like they're calling sysctl to grab > just > CPU or just memory, etc., metrics. > > On the InfluxDB side of things, I'm not seeing any Telegraf plugin to > monitor > sysctl in a generic way, although you could probably emit InfluxDB line > protocol pretty easily from the exporter tool or something similar to that. > > My impression, from a monitoring perspective, is that there isn't much > demand > to gather _everything_ in sysctl in a single call, although categorized > subsets > might be appealing. Many sysctl nodes have a) opaque binary values, w/ version depended formats b) opaque binary values w/ size more then values c) very large size: sysctl -b kern.msgbuf|wc 10200 55998 1048505 Also, InfluxDB is hipster shit and for many cases need to preprocess collected values, because imposible to calculate on Influx side.