Monitor the customer, Feed the Business

Slide 1

Slide 1 text

Monitor the customer. Feed the business. Ben Whaley @iAmTheWhaley

Slide 2

Slide 2 text

We are not focused.

Slide 3

Slide 3 text

What's missing from this picture?

Slide 4

Slide 4 text

The customer The business is not represented.

Slide 5

Slide 5 text

Notifications

Slide 6

Slide 6 text

Customer traffic 50% of normal. Signups doubled in the past 15 minutes. Mobile app crashes for 9% of users immediately following a release. The current notification strategy is to hear about it from the users on twitter. This is bad. Notifications

Slide 7

Slide 7 text

Capacity

Slide 8

Slide 8 text

Visualization

Slide 9

Slide 9 text

TCP Connections in CLOSE_WAIT Restart Restart Restart ... Defects Asleep at the keyboard

Slide 10

Slide 10 text

must be collected, measured and acted upon. All this Data

Slide 11

Slide 11 text

collectd nrpe Metrics sys logs app logs snmp jmx (net|io|mp)stat, sar nmap nessus munin jvm network stack Collect Measure graphite uptime riemann Icinga/Nagios Logstash Sumo Logic Loggly Splunk Cacti RRDtool ganglia Datadog Boundary tasseo

Slide 12

Slide 12 text

If you cannot measure, it you cannot improve it. Sir William Thompson, AKA Lord Kelvin cartoon via toothpastefordinner.com

Slide 13

Slide 13 text

Act?

Slide 14

Slide 14 text

Alignment. Act.

Slide 15

Slide 15 text

Stakeholder support. Act.

Slide 16

Slide 16 text

Share. Act.

Slide 17

Slide 17 text

Automate. Act.

Slide 18

Slide 18 text

Alignment. Stakeholder support. Share. Automate. Act.

Slide 19

Slide 19 text

Business

Slide 20

Slide 20 text

Monitor the customer. Feed the business. Ben Whaley @iAmTheWhaley