Is It Time To Version Observability? (Signs Point To Yes) • Charity Majors • GOTO 2024

This presentation was recorded at GOTO Amsterdam 2024. #GOTOcon #GOTOams

RESOURCES

ABSTRACT
The DevOps revolution is winding down, and we’re entering the post-DevOps era. We’re at the precipice of a massive generational shift in how we build and understand our software, and engineering teams need to prepare.

In the past, telemetry focused on basic metrics: reliability, uptime, MTTR, MTTD. Observability 1.0. Companies that settle for these basic data points will struggle to keep up with modern development patterns, while hemorrhaging money.

As engineering best practices around separating deploys from releases, testing in production, and observability-driven development have gone mainstream, the metrics-driven approach to telemetry has stalled, and it’s time for a new version: Observability 2.0. Learn what this new version means for your engineers, and how to embrace this breaking change to:

• Save them from drowning in symptom-based alerting
• Help fewer people work together to build better software
• Create fast feedback loops throughout the entire organization through highly granular visibility into all their systems [...]

TIMECODES
00:00 Intro
00:47 Every company is now a tech company
04:57 Fast feedback loops
07:46 Brief history of observability
09:54 Observability 2.0
14:48 Meh-trics
17:36 Who uses it & how?
19:30 How you interact with production
20:44 How you debug
23:00 The cost model
30:07 What to do to move towards Observability 2.0
32:01 What got us here won't get us there
35:05 Outro

Download slides and read the full abstract here:

RECOMMENDED BOOKS

#Observability #DevOps #PostDevOps #Reliability #MTTR #MTTD #ObservabilityEngineering #CharityMajors

CHANNEL MEMBERSHIP BONUS
Join this channel to get early access to videos & other perks:

Looking for a unique learning experience?

SUBSCRIBE TO OUR CHANNEL - new videos posted almost daily.

[kashc]

Canonical logs, such a powerful pattern.

Obs 1.0: we need to slog for structured logs..
Obs 2.0: who let the logs out!!?!

[brnto]

10:38 logs are just bits of data that haven’t been mapped into your data model. If it’s of ongoing interest, it should probably be persisted in a database with referential links to the relevant database records. Then you can properly query the data and link it to all the relevant information.

[jscancella]

Interesting talk. Too much time spent trying to convince, instead of showing examples of why canonical logs are better and how to easily do it and other tools that support them. Hopefully since openTelemetry still hasn't done much in the logging area, they can pull these ideas and make it simple to implement

[Mindjux]

Show request.latency in wide structured logs please, and calculate the price as well