This is a second blog post in a related series, the first is located here.
Using SaltStack to run distributed monitoring checks is fun, but what is really awesome is writing the result of all those checks to CouchDB. Instead of returning the result of the check to the command line, or to the master even via programmatic method, you can use SaltStack’s returners to have the output returned somewhere else. There is a CouchDB returner which works very nicely for storing the result of nagios and other network checks (network.traceroute for example).
What I’m hoping to eventually end up with is a hybrid monitoring/alerting solution, which will provide more detailed alerts than a simple “host down”, or “service down/timed out”, something like Blog Post on Sys Advent, but instead of showing history from one nagios instance, providing a short historic, global view from from various DigitalOcean and Vultr VPS’s. Informative stuff like “Host (name) down, 4 of 7 locations reporting down at (timestamp)".
Over the holiday break I am going to invest more time in copying my nagios monitoring setup/configs to SaltStack Pillars, and begin to run all my checks globally, using
cron at various intervals for different Pillars. I will then need to write some code to:
Once I get these steps completed, I will work on packaging the resulting scripts into something which others can perhaps find useful as well. If I come up with a good name for this conglomeration, I’ll register a domain for it, and start an organization on Github for it as well.