The case for Monitoring

(Food for thought)

Nir Cohen @ Gigaspaces

nir0s @ GitHub

@thinkops

Agenda

• Cultural Aspects
• The case for Open-Source
• Enabling Automation

Cultural Aspects in Monitoring

Everyone owns monitoring

(not just Ops)

Discuss system state periodically and adjust

(not just once)

Monitoring should be part of the application

(not an after-thought)

Work towards prevention

(not fire-hosing)

Be open to change

(otherwise you can't progress)

Be open to open-source

(free isn't necessarily bad)

The case for open-source

Huge ecosystem, loads of tools

On-going support by the community

More transparent future

Contributing back

Monitoring Enables Automation

Continuous Deployment

Auto-scaling

Auto-healing

After-thought

Monitoring Continuously Evolves

Monitoring systems must be more available and scalable than the systems being monitored

​Monitoring is hard. You can simplify it only so much before its value decreases.

Learn!

• Continuous Delusions (Avishai Ish-Shalom@Velocity) - https://goo.gl/5nlSze
• Why Percentiles don't work the way you think - https://goo.gl/8tQWB7
• InfluxData - https://goo.gl/EAZtky
• The Math Behind Bad Behavior - https://goo.gl/5h39we
• Stream Processing... - http://goo.gl/5KXNc6
• Incuriousity will kill your infrastructure - http://goo.gl/1ooJaN
• Practical Guide to StatsD/Graphite Monitoring - https://goo.gl/CHkAcu
• How we scaled our monitoring platform - https://goo.gl/XFWB3P
• 5 years of metrics and monitoring - http://goo.gl/WXbgxh
• Please, no more minutes, milliseconds, monoliths or monitoring tools - http://goo.gl/A9AZa2
• StatsCraft 2015 Sessions - https://goo.gl/i0nZTb
• Monitorama Sessions - https://goo.gl/0Hl8pe