Logs, metrics and traces have come to be known as the three pillars of observability. Now, there is a push to have a fourth pillar — snapshots. In this four-part microwebinar series now available on demand, we examine the observability and discuss how snapshots make it easier for developers to troubleshoot issues. Joining SD Times is Liran Haimovitch, co-founder and CTO at Rookout, which was recently acquired by APM and observability platform provider Dynatrace.


Episode 1: The Fourth Pillar of Observability

Why should SREs get all the fun? Everyone knows logging sucks and so SREs created two new pillars: metrics and traces to make their jobs easier. Well, logging sucks for developers too and they deserve – and need – better Observability. That’s where the fourth pillar of Observability comes in: Snapshots. 

  1. How logs are very low fidelity – How snapshots are high fidelity
  2. How much tinkering is required to optimize logs
  3. Snapshots are allowing you to see the data in the context of the code while logs require you to do a lot of back and forth
  4. How Snapshots are highly reliable while logs make a lot of assumptions and can be wrong and outdated
  5. How one of the best things about Snapshots is how you can add them in real-time without requiring code changes or deployments. More on that in the next episode.

Related Resources: 



Episode 2: Live Observability 

In the first episode, we will discuss the fourth pillar of Observability and how Snapshots are so much better than logs. One of the big benefits of Snapshots in particular and agile Observability in general is that you can adapt your Observability in real-time without requiring code changes or redeployments. 

Related Resources: 



Episode 3: Cut Observability Costs with Developer-First Observability

If you dig down to the bottom of it, you’ll find that Observability will eat up any budget allocated to it and then some. That’s because the need for more Observability is rarely rooted in engineering needs. It is in fact coming from a much more primal place: the fear of the unknown. This is why Observability is a huge cost driver, growing year over year, and tremendously hard to optimize. Because anything you cut away will be quickly replaced by new data points. 

The irony is that over 95% of the data is never used. Engineers are collecting and storing it out of FOMO – the fear of missing out on logs and other Observability pillars that may be useful one day. Dynamic Observability is how you can turn it around on its head.

Related Resources: 



Episode 4: Troubleshooting & debugging across the SDLC 

Everybody is talking about the importance of shift-left and they’re definitely right. But – little by little, we’re seeing the opposite movement. We’re seeing the SDLC itself shifting right. Today, engineers spend most of their time developing and testing code in production-like environments. To meet those challenges, we need to adapt appropriate Observability tools throughout the SDLC.  


DMCA.com Protection Status