 I just have to take another picture. Perhaps raise your hands. Yoo-hoo! Speaker's perspective, thank you. Same for the introduction. Like you mentioned, my name is Banda Erz. That's a little bit stupid in English. I'm working at NetRays in Germany and I'm one of the co-founders of the I thinker project. I started a little bit in the reporting area there and now my responsibilities are more project management, community work, conference talks and stuff like that. But I see that the slides are a little bit screwed up. Okay. So, about I thinker, how many of you know I thinker or have heard of it? Or kind of? Nadjo's? Okay, I think I had the roots in Nagus, but it's much, much better. So even on Nagus, you definitely will love I thinker. I can say for now. I thinker is an open source monitoring solution. It focuses on monitoring, on alerting, on gathering BI data for performance, for capacity planning and stuff like that. So originally we forked I thinker out of Nagus, version 3 in 2009. We take the code-pays from there, but since the couple of years, we totally developed our independent core, which is I thinker 2, which is production ready since June last year. It's a totally different code-pays and independent from the old Nagus work. We are a group about 27, 26, developers depends on the free time working on I thinker. We started primarily in Europe, but now we are well-distributed out of the US and South America. And we have different teams in the project from day one, means we have independent responsibilities. Every team, the web, the core, they handle their stuff on their own and they are independent from the others that we are quick. The tool stack we have is we have the I thinker core. It's in the middle, it's probably the old solution. We still support, we do bug fixing, there will be a new release coming out tomorrow of the old core as well. We have a reporting framework based on Chester reports. We have a couple of web interfaces. So we have the old classic interface, which is a beautified Nagus interface, and the Nagus interface, it's hard to beautify, but you can do a little bit with images and styles and stuff like that. Then we started to develop something new, it's a singer web based on XJS, but now we decided as well to do something new, which is singer web 2, which I can show you later. But today I will focus on that I thinker 2 and I thinker web 2. A short introduction if you have no clue what I thinker is. Probably there are three main responsibilities for the monitoring system. The first of all is monitoring stuff that can be hosts, computers, storage, switches, or it can be doors. If you want to monitor if a door is closed or somebody opens it and you can do it, we have tools and there is hardware you can monitor as an MP. So you probably can monitor everything. So that's one thing, getting information about hardware and one important thing in our singers that the primary approach is doing monitoring in a regular interval every minute, every five minutes checking the system. So really actively checking a system, not waiting for an agent bringing data back to the monitoring server. If something is wrong and normally every time something is wrong, then you do a reaction of that. So you can execute event handler, so you can send out notifications, typical email, short messages, you can do voice alerts, meaning you can do text to speech and you can listen to the error message on the phone and dial a code for acceptance. You can open up a ticket or something else, however whatever is needed to keep somebody informed about that. Another thing is like bringing the data out of iSinger because on every monitoring check you get a lot of performance data so you can see how fast as a web server response is in SSL certificate, still valid, and you can bring the data out. So we have standard adapters for Rockstash, for Greylock, you can directly write to Graphite, you can write to OpenTSDB since version 2.3 from iSinger which is out since Tuesday, you can write to InfluxDB so these time series databases are supported natively so you can directly write to them. You can generate performance data so you can do capacity management, you can do a little bit of prediction but prediction is very hard. It's easy looking at a disk, you can see the disk could be fallen two or three months, but doing capacity management and forecasting for load is not so easy but you can do. And as well you can do reporting based on the data because you can write them to the database and this is a pretty easy source for reporting as well. So what it is iSinger 2 is based on C++ the other single core is just C but it's not scaled very well you can make a single CPU hold on a good server but don't use the others. We have a powerful CLI to support iSQL and Postgres databases you don't need to run the database for the monitoring system but you can do and if you want to do so we have to decide between all of them. And we have background available so you can start it very easy just clone the git fire up the background environment and you can see how the cluster works or a single instance iSinger works. So if you are in that area of configuration and if you know how a Nagis configuration or iSinger 1 configuration look you have to run something because we broke up with the configuration it changed a little bit but just to give you an idea what we can do now with iSinger 2 is we can have something like rule-based applies so for example we can based on host variables imagine like facts you have in Puppet you can apply services to specific hosts in the example number 1 or like conditional so we support if else if rules where we can set specific monitoring thresholds depending on time zone for example imagine that you have different thresholds during the normal day shift and what react if your disk gets full and about 60% but it don't once bring somebody out for 60% disk at night so we can do something like time limit thresholds with a pretty cool feature came with 2.3 so the architecture is that we have the iSinger core and we have a couple of modules supporting different features we have so writing to database writing to graphite supporting cluster stack and you can enable or disable every one of these features it's done here the iSinger C like command you can enable the feature or disable it it works primarily like a patchy module or patchy website it's creating a soft link of that feature means that you can have multiple iSinger instances working together I show you that in the cluster slide and you can decide on which node you want to write to the database you want to write to graphite or execute active checks so the cluster a big challenge in the old monitoring partners was distributed data so e-replicated data also means from the country also from the security perspective so specific network zones where you're only able to connect into or receive data this was possible in the old world like transferring the files or using SSH but now we have a cluster stack based on SSL and certificates you can use an existing certificate infrastructure or if you still have one for puppet you can use them as well for iSinger and the cluster stack is included means you can create mesh up clusters you can have multiple iSinger nodes with different responsibilities and if they have the same knowledge means the same zone then they share the data means for example the server in the left upper corner can execute a check but you can write the result to the database on another server in the cluster or somewhere else so all the servers share the data and if you want to limit specific knowledge to a specific area you can create something like a zone so if you have only for example a specific country you can have two monitors and there are only two servers which are able to reach the servers there then you can create a zone you can create multiple zones and they are syncing the state they are syncing the configuration files as well so you don't have to take care about syncing manually the files, everything is done wire a single port and a secured single port so the web interface iSinger web 2 it's not finished it's not this week, hopefully tomorrow I'm not in the office but I hope they will do their work it's hopefully be on the website tomorrow the beta 3 version it's based on PHP we're supporting multiple backends so we can read out from the database you can read out from live status it's very fast and we also took care about all the security layers needed iSinger web 2, imagine it the basic framework the basic framework of the iSinger web 2 interface just takes care about the backends, so reading writing from the databases takes care about identification you can use internal users you can use a directory and takes care of the installation but it's just the foundation means you can imagine the framework like something you have with perhaps WordPress and a plugin iSinger web framework has no idea about monitoring it just provides the basis for the user and then monitoring is a module for that and monitoring is a module we deliver because without monitoring it doesn't make sense to enter the framework perhaps because it's a monitoring solution so it's in there, we ship it but you can add more so we have a document model we have a business process model the visualized business process you can show up the web interface and also P&P for Nagios it's like an R&D add-on for Nagios iSinger where you can write your performance data in R&D so how does it look like? this is the landing page so if you know the old interfaces where we have the number crunching in the top bar we reduce it, we kick them out because it doesn't make any sense to see that you have 3000 hosts which are okay it's nice but it doesn't make sense so we focus more on the problems like on the left side you see the current service problems or on the right side the recently recovered problems so if you were out for lunch you came back you see what's happened, so what's okay now but we have a problem perhaps half an hour ago everything has a column design so if you click on a row here, I hope the laser pointer works oh yes welcome to the 20th century you see the details on the right side so if you click a host you see all the details you can, for example recheck reschedule a check for the host with one click so we don't have to go to reschedule the SI1, do you really want it now so it's just a single click to execute a recheck and then one of the last you have possibility to filter in detail so you can filter every state by name or whatever you want to do like you see in this example it's a little bit small, but I hope you can read it then you can store these filters and add specific filters to the dashboard like filtering for a specific server type specific state, responsibilities host groups this is pretty easy and also in the host groups overview where you have more information in a much, much smaller design it's also responsive interface so if you use this on an iPad or an iPhone just one, three, if you have a big screen then it could be that you have five or six colors so it depends on your screen size so what we have for now we don't know who our users are because they don't have to sign up they can just use the RPM or packaging whatever, so we don't know there are some customers we know that they are using it massively we see a lot of people coming over to sing a tune now directly we do a couple of community events our next overview in Asia will be in Kuala Lumpur in June, if you're interested we are searching for speakers and probably we're searching for people coming to the camp, that would be great and then we do one in Portland so if you jump up at Kornf it's the day after Kornf in Portland, Oregon come to end try it out, if you know background close the gate start up the background box rethink your configuration so if you have an existing monitoring system based we think about the configuration try the new features we bring with the single tune play with the web interface it would be great to give us feedback if you like it or we should improve and that's it, thank you very much and hopefully how to compare the sensor, it's different make it quick or it's hard to explain I think sensor has a totally different stack I think the basic stack of sensor is more complex sensor has a different approach and what I see but this is just my personal opinion that sensor goes enterprise so there's an enterprise version of sensor and normally nothing against an enterprise version it's another different model but I think that all the new and funny things will land in the enterprise version which is not interesting for me and this is what I don't like I hope that sensor is a good open source monitoring alternative but seeing the new features going enterprise is something I don't like Thanks, catch Bernd when he gets on stage as well I think he can answer more in person thank you so much Bernd, big hands for Bernd