Article Blog Image

Webinar: Effective SLOs

Events

Let’s get real: Service Level Objectives are hard to get right. They are indeed a transformative technique in making services reliable, however there are many potential pitfalls and antipatterns when implementing them that can lead to frustration. Let’s explore several that I’ve observed in my career!

Common SLO Pitfalls

SLOs can be hard to explain

It can be a challenge to clearly articulate what they are and the value they provide, and quoting the SRE Book isn’t going to be enough for non-technical members of your team or your stakeholders.

SLO can mean completely different things depending on who you speak to. For example, I offer encountered the conflation of “Service Level Objective” with “Service Level Agreement” in conversation with internal customers.

How can you develop a healthy and productive SLO process when there isn’t agreement on the definition and purpose of the term?

Essential observability and technology are missing

I remember onboarding to a team that said they had SLOs in place, but there were no metrics and certainly no dashboard to visualize/monitor their performance! This made the SLOs completely useless as there was no method to verify they were being met.

Yes, this was an extreme case, however, the point is made: SLO can’t be implemented without gathering the right telemetry, storing it, and being able to query it. In many cases, you will need to start from scratch.

SLOs are defined too loosely to be useful

When SLOs are too loose, the engineering team won’t get paged… until customers complain and escalate to them anyway!

When they are approached as a checkbox rather than a way to improve system reliability, SLOs can become detrimental and lull the team into a false sense of security!

Violations aren’t enforced

I’ve seen several instances where SLOs are being violated but the engineering team continues to ship features.

Usually there is some overriding factor: a deadline, some important feature, or inconsistent messaging from management that prevents the team from focusing on reliability despite all the warnings.

Introducing ‘Effective SLOs’

Is the goal of having well-implemented SLOs that provide tangible business value merely a pipedream and therefore a waste of time?

No.

As someone who has successfully rolled out SLO across many teams, let me assure you: they are within your reach- so long as you approach the task in an intentional manner!

I will be presenting the free webinar “Effective SLOs” on June 13 at 2PM ET, where I will provide a step-by-step approach to implementing Service Level Objectives while avoiding the types of pitfalls I described above. This content is based on almost a decade of experience in Site Reliability Engineering across medium and large tech companies.

You can sign up here.

Again, the webinar is free!

Tags: