Having difficulties with passively testing your infrastructure with old-school supervision software ? Your infrastructure description differs from your monitoring configuration ? Having trouble communicating between ops and monitoring ? Use Salt configuration management to change that.
Once you start using configuration management systems like Salt and describe your infrastructure as code, you can then re-use that description to generate monitoring configuration for various existing monitoring solutions (nagios, shinken, etc.).
You can also use Salt to orchestrate the checks and use other data silos to explore your supervision.
Active checks can be automated by Salt by reusing nagios scripts and munin plugins, collect the data in Graphite and explore them by building graphs in Grafana, finally define alerts with tools like graphite-beacon or Cabot.
Adopt Salt and its flexibility, modularity and event bus that will add software glue between your existing applications and help you scale them.
This presentation will be looking at how to solve some scalability issues in monitoring and migrating from traditional tools to newer tools while keeping existing developments.
Open Source Stack used:
Speakers: Arthur Lutz