Our Marco Fretz has visited the OSMC (Open Source Monitoring Conference) 2018 in Nuremberg and reports his impressions below.
The OSMC 2018
The four-day OSMC with around 300 visitors is all about monitoring tools, concepts and the automation of the monitoring system. Of course, Icinga2 is very well represented – in the talks, among the participants and organizers. The conference is organized by NETWAYS, the company behind Icinga2, and takes place annually at the Holiday Inn Hotel in Nuremberg.
On the first day there were workshops on various topics and on the fourth day a hackathon, with projects that were determined spontaneously. The English and German talks took place on the second and third day and were divided into three tracks. Thanks to the agenda (online, in front of the rooms and at the back of the badge), it was usually easy to decide on the most exciting track. Talks that you missed can be watched later online as video.
I only visited days two and three again (with the talks). With around EUR 1500.- for two days conference, three evening events and three nights in the hotel, the conference is also very attractively priced.
In any case, the OSMC is also known for the rich and excellent catering (all-you-can-eat) from breakfast to the evening events and the smooth organization – the hotel and conference check-in are completed in under a minute. This has come true again this year.
More impressions: #OSMC
For me, the most important thing was to meet new and familiar faces and the related exchange about their own and other monitoring landscapes including their problems and solutions as well as, of course, the direct line to NETWAYS / Icinga. Thanks to the talk from @ekeih (scaling Icinga2 …), some people quickly found themselves running Icinga2 in a setup similar to ours.
Exciting talks about Prometheus or concepts that rely on Prometheus made me feel that Prometheus is becoming more widespread, not only in the “cloud native” world but, for example also for HTTP SLA monitoring (MORITZ TANZER, nic.at), network monitoring (MATTHIAS GALLINGER, ConSol), etc.
Our current plans at VSHN for the integration of Prometheus into our monitoring environment were thus confirmed.
A big bottleneck in Icinga2 and Icingaweb2 is the IDO database (MySQL / Postgres), whose schema dates back to the Nagios and Icinga1 era and has been steadily worsening over time. At that time, a relational database for actually volatile status information such as service and host states, etc. seemed to make sense. In larger setups, however, the writes from Icinga2 to the DB are the bottleneck. Also, the query performance of Icingaweb2 suffers greatly in certain configurations for larger setups.
A lot of details are not yet known, however, Redis is used for the volatile status information and a SQL DB for the historical data. A first version already runs on a trial basis at Icinga and was presented in a live demo. I think it’s great that you can probably use the IDO DB and the IcingaDB module in parallel (transitionally), the same applies to the Icingaweb2 monitoring module – this will greatly simplify migration.
To take away
Try it out…
OpenAPM is not a tool in itself but shows in a simple way which tools can be combined with each other to build an application performance management / monitoring landscape. Just try it here: https://openapm.io/landscape
Certainly exciting are the maps for Icinga2. You can give each host or service object the geolocation via custom variable, then they are automatically displayed on the map and grouped according to the zoom factor: https://github.com/nbuchwitz/icingaweb2-module-map
From the talk of @ClaudioKuenzler a plugin for easy monitoring of Rancher2 and Kubernetes: https://github.com/Napsty/check_rancher2
I have taken great ideas with me, met new people, ate a lot (and yes, the gin was good too ). A great conference that’s really worth it. I like to go again.