Technology changes and process changes in how people build and manage Internet systems have driven a need for a new approach to monitoring. We talk about why, what and how.
1. Welcome to the webinar! We’ll start in just a few minutes.
Please submit all questions in the Chat panel. We’ll cover them at the end of the webinar.
Theo Schlossnagle - @postwait
Founder @Circonus
Monitoring the #DevOps Way (and the future of monitoring)
3. DevOps
So many definitions of DevOps…
Applying software engineering practices to improve
operations and infrastructure management.
Spreading operational acumen through the organization
(specifically into product engineering)
https://www.flickr.com/photos/jamieanne/5744219523
4. The days of “is it up” are long over.
Monitoring has changed Significant challenges in
operability and accessibility
5. The days of “is it up” are long over.
Monitoring has changed Significant challenges in
operability and accessibility
6. The days of “is it up” are long over.
Monitoring has changed Significant challenges in
operability and accessibility
7. The days of “is it up” are long over.
Monitoring has changed Significant challenges in
operability and accessibility
8. The days of “is it up” are long over.
Monitoring has changed Significant challenges in
operability and accessibility
9. Managing risk and exploiting opportunity
Operability
Monitoring (data collection and analysis) is hard.
Now add “always available” and “always immediate.”
Many companies squander opportunity in engineering by
building monitoring solutions instead of more direct
market value.
ALWAYS ON AND ALWAYS WORKING…
EVEN WHEN YOU’RE INFRASTRUCTURE ISN’T
https://www.flickr.com/photos/coastguardnews/2291992191
https://www.flickr.com/photos/ussnewyork/5403986347
10. Embracing today’s engineering organizations
Accessibility
If it isn’t accessible via an API, it isn’t accessible.
Configuration, analysis, alerting, visualization…
it all must be accessible or it simply won’t integrate.
This is how engineers work.
The dynamic nature of the the cloud forces your hand.
used with permission from Stacey Mason @stcymsn
11. Looking for information in the data
Millions of streams
Is simply too much to look at
We now must leveraging mathematics
(e.g. statistics and machine learning)
to surface interesting information
https://www.flickr.com/photos/vickyb/493019855
14. Solving this at scale
Once you solve the scale problem,
what are the opportunities?
1000 measurements / second?
How about 1 million? or 1 billion?
Imagine the insight.
Now imagine the data science problems.