<p>(Sorry for the late reply; somehow this thread got lost in the mess)</p> <p>On Oct 12, 2011 2:03 AM, "James" <<a href="mailto:wireless@tampabay.rr.com">wireless@tampabay.rr.com</a>> wrote:<br> ><br> > Pandu Poluan <pandu <at> <a href="http://poluan.info">poluan.info</a>> writes:<br> ><br> ><br> > > The head honcho of my company just asked me to "plan for migration of<br> > > X into the cloud" (where "X" is the online trading server that our<br> > > investors used).<br> ><br> > This is a single server or many at different locations.<br> > If a WAN monitoring is what you are after, along with individual<br> > server resources, you have many choices.<br> ></p> <p>It's a single server that's part of a three-server system. The server needs to communicate with its 2 cohorts continuously, so I have to provision enough backhaul bandwidth from the cloud to my data center.</p> <p>In addition to provisioning enough RAM and CPU, of course.</p> <p>> > Now, I need to monitor how much RAM is used throughout the day by X,<br> > > also how much bandwidth gets eaten by X throughout the day.<br> ><br> > Most of the packages monitor ram as well as other resource utilization<br> > of the servers, firewall, routers and other SNMP devices in your network.<br> > some experimentation may be warranted to find what your team likes best.<br> ></p> <p>Currently I've settled on a simple solution: run dstat[1] with nohup 30 minutes before 1st trading session, stop it 30 minutes after 2nd trading session, and send the CSV record via email. Less intrusion into the system (which the Systems guys rightly have reservations of).</p> <p>> > What tools do you recommend?<br> ><br> > OH boy. I like JFFNMS very very much. It has a very old version in portage<br> > (masked) but a very new version out there for Debian and Ubuntu. It<br> > runs on all nix, if you want to driectly compile and install.<br> ><br> > I'll be putting together a new ebuild, as soon as I get it working<br> > with the latest postgresql. Mysql works out of the box. Postgresql-9<br> > has many new and very cool features.<br> ></p> <p>Cool! I *love* Postgresql! Update me when the ebuild's done?</p> <p>> > Remember: The data will be used for 'post-mortem' analysis, so I don't<br> > > need any fancy schmancy presentation. Just raw data, taken every N<br> > > seconds.<br> ><br> > Personally, I have some large, high risk design work going on. JFFNMS<br> > and pg9 are the best choices from my research. A whiz like yourself<br> > could easily look at the old JFFNMS ebuild and create a new one.</p> <p>Naaah, I'm going to wait for your ebuild. I'm sometimes lazy, you know ;-) </p> <p>> PG-9 (please no flame wars on mysql vs pg9) is very cool and what<br> > my work is migrating too, once I get some breathing room.<br> ><br> > Craig at <a href="http://jffnms.org">jffnms.org</a> is very cool and responsive. He also works closely<br> > with those that submit patches. Nagios is a large, disorder array that<br> > had many devs fork off since the project leader (was/is an a_ole)<br> > is quite difficult to work with.<br> ></p> <p>That sounds really cool. I've been hesitant to go the Nagios route because of the mess. I'll sure to be checking out JFFNMS.</p> <p>> JFFNMS rules and is very cool for managing cisco and other routers,<br> > not to mention a myriad of snmp(1,2.3) devices and all types<br> > of servers. The original guy, Javier, was snapped up by someone<br> > worth billions, to manage and extend his financial network, but, Craig<br> > is probably stronger coder, and extraordinarily nice human being.<br> > It's mostly php. Lots of folks extend JFFNMS, Craig keeps it clean<br> > and well written and documented code.<br> ><br> > <a href="http://www.jffnms.org/">http://www.jffnms.org/</a><br> ><br> > hth,<br> > James<br> ></p> <p>Thanks for the heads-up! Although the original problem is solved already (granted, in a somewhat kludgy way), your post is a great write-opener! Much appreciated :-) </p> <p>Rgds,<br> </p>