A DAY in the LIFE of a DATA CENTRE | HOW WE MONITOR | EP 1 | BRIEF NOC OVERVIEW!

[Applause] so guys welcome back to another day in the life which is kind of an impromptu you won I've just been installing some videos for a new rap setup and part of the setup process I go through our plugins and fans into them at different settings so that they draw different current I thought you know what it's not a chat with marketing guys we do a video on our monitoring system and what software we use and how we make it all work so I'm going to go get the fans plug them in ramp up two of them to different levels get some readings we'll go check out the systems roll the introduction [Music] so you take these fans in set one to low heat one to high heat one will pull four amps want to pull a temps which means I can get reading on the graphs and make sure that everything is monitoring correctly so I've got them set to heat and one of them is set to one and the other one set to 2 so the one that's on one good tool for it's like 4.2 amp so I'll flip it into PD and just make sure it actually works and then I'll plug the other one into 3p also make sure it works so they're spinning up there on and they're hot I have to leave it a couple of minutes for the software to actually call it and take a reading and then as you can see here feed a which is what if like the low powered one into is pulling 3.92 m and fifi is pulling 7.6 for them as you can see here it's got it like so so guys in here we are in the NOC and these are the monitoring systems that myself and the guys monitor on a daily basis so I'll give you a quick run-through of what we've got going on up here and explain some of the software we're using and why we use it so on these two screens we monitor power calling and network this is running as its that screen and one of those screens it's running a piece of software called cacti it's an open-source bit of software we've taken it and we've heavily customised it to do a lot of back-end stuff as well so these are just the snapshot graphs that we review in the NOC just a quick look up from you computer just to kind of see what's going on on the computers we do view in hell of a lot more detail as well right down to the T so we monitor 20,000 data points at this moment in time obviously that's ever-changing as we add new temperature sensors new servers whatever it is we monitor it obviously it's all done with SNMP so that's kind of cacti we'll go over in a bit more detail in one of the other videos kind of how it works how to install it what it looks like when you're actually on it so on this top middle screen you might have seen in our gentes video that we done it's going to be blurred out because there's a lot sensitive information up there but we've effectively got top left we monitor latency to various targets so that could be transit services that we have we checked the latency across those and it starts banging and flashing when it all goes wrong we've also got we've got some custom software which is used for dos attack detection so we monitor various parameters on our network that help us work out whether there's an attack or whether it's just a genuine flow of traffic and we've also got the generator stuff which you listen in the gen test video tells us the state so generators what their loads are etc so if we wander down this side a little bit so we've got a snapshot of our CCTV on this screen it's literally just an overview of probably the more sensitive areas and it scrolls through different screens things like doorways in in and out of the data center air handler plants you know build room this NOC etc it's just an overview we do have a lot more cameras on our computers that we can see exactly where they are what they're doing who's where and on the screen above this one is a syslog screen it literally just spits out syslog events from our devices and this is also running in cacti which is what the other screens are running on for the graphs and weather maps and stuff like that we've also got a nice little bit of software up here that we've written ourselves it talks about access control pulls out people that are signed into the datacenter tells us where they lost what was who they are what company they work for what what they actually doing around site it's also hooked up so that in the event of a fire that prints itself out so we've got a checklist of who's on site so we can record everything that way there is there are some other bits of software we use observe um it's another bit of software we use again it's open source there is a pay version which obviously we have and it's just good to have a few different monitoring systems monitoring the same things because they work in different ways some are more simple whereas like observing them is very simple it's just graphs and it tells you what you need to know cacti really fine details you could do a lot with it it's very powerful yeah that's pretty much all there is to it software wise we do have a few more bits of software but I'm not going to keep rabbiting on it's just kind of an overview this video so Pavley is going to be doing a video with you guys soon about the physical side of it so the sensors the servers that make up the monitoring platforms the custom stuff we've done to monitor various bits and bobs so Pat will take you on a quick run-through of that in the upcoming upcoming videos we've also got James who's going to go through kind of a lab set up with you so installing it playing with it and just setting up some devices on say cacti for instance I think that's what it's going to use so yeah we're gonna wrap up this video because I've got some stuff to crack on with I've got some very urgent setups to process so remember to leave a like comment with any suggestions you know we love doing videos that you guys want to see this being one of them and just share the video spread spread the love can we see you all in the next show

Channel: Custodian Data Centres

Views: 63,465

Rating: 4.9474673 out of 5

Keywords: datacentre, NOC, DCIM, cacti, monitoring, NOC engineer, cloud, networks

Id: 12WPgZTPtx0

Length: 6min 50sec (410 seconds)

Published: Thu Feb 13 2020