Monitorama EU came to Berlin - 2 days full of interesting talks and workshops on the field of monitoring and open source. All this took place in the
Urania.
For me, the most notable talks on the first day were the two keynotes by
Dylan Richard and by
Danese Cooper and the talks held by
Katherine Daniels and by
Ryan Smith.
Dylan worked behind the scenes of the most recent Obama campaign and gave an insight into their monitoring strategy and their scalability challenges, whilst Danese shared some of her experiences as an Open Source Evangelist.
|
From Danese Cooper's presentation: propaganda against open source |
Katherine's talk was about frustration concerning monitoring in classical Ops teams and about how to improve the situation by reducing complexity and re-focusing. Ryan's very entertaining talk took the weaknesses of monitoring systems in aircrafts as an example and explained how to deal with those weaknesses and what to learn from that.
Day 2 consisted of two parallel sessions: one for talks and one for workshops. This gave a good mixture of theoretical and practical information.
The day started for me with a talk by
Reza Spagnolo - a colleague of mine - about adaptive application architecture. The talks by
Jeff Weinstein and by
David Goodlad were also very good. Jeff gave a good overview of the different roles in a company and their expectations concerning metric data. And David emphasized that business metrics should be designed before system metrics and that alarms should be based on business metrics.
The workshops I attended were about
Kale (by
Abe Stanway), big
graphite installations (by
Devdas Bhagat) and graph automation based on tools like
collectd,
statsd and graphite (by
Michael Gorsuch) - all in all very nice and definitely worth the time spent.