veyepar: Rethinking monitoring with Prometheus

Author(s):	Martín Ferrari
Location	Heidelberg
Date	aug Thu 20	Days Raw Files
Start	11:00	First Raw Start	error-in-template
Duration	00:45:00	Offset	None
End	11:45	Last Raw End
Chapters
Total cuts_time	None min.

https://summit.debconf.org/debconf15/meeting/312/rethinking-monitoring-with-prometheus/
raw-playlist raw-mp4-playlist encoded-files-playlist public mp4 svg png

assets release.pdf Rethinking_monitoring_with_Prometheus.json
logs
Admin: episode episode list cut list raw files day marks day marks day image_files

State:

Locked:

clear this to unlock

Locked by:

user/process that locked.

Start:

initially scheduled time from master, adjusted to match reality

Duration:

length in hh:mm:ss

Name:

Video Title (shows in video search results)

Emails:

email(s) of the presenter(s)

Released:

has someone authorised pubication

Normalise:

Channelcopy:

m=mono, 01=copy left to right, 10=right to left, 00=ignore.

Thumbnail:

filename.png

Description:

markdown

This talk will introduce a new tool for the discerning sysadmin: Prometheus.

It will start with an overview of the challenges of packaging a Go application in Debian: the small number of packaged libraries, the very particular model of dependency management and compilation, etc. It will be followed with an overview of the system and a taste of what you can do with it. If time permits, I might even do a small demo.

Most sysadmins face a similar problem: the state of FOSS monitoring tools is still stuck in the 90s. Nagios, Munin and friends get the job done, and you might think this is enough. But once you have a taste of what you can do with a tool like Prometheus, you will never want to come back!

Prometheus is a very different system. Instead of checking good/bad states of services, or just nicely graphing values, it collects information. A lot of information! At its core, it is just a time-series database and data collector. What makes it powerful is the advanced calculations you can make on the fly with the collected data, from which you can derive nice graphs, fire alerts, or just store for future analysis.

Comment:

production notes