Much That We Have Gotten Wrong About SRE
An illustrated summary of Developers ➡ DevOps ➡ SRE
Much That We Have Gotten Wrong About SRE

1. Developers wanted to ship their produce

To the other side

2. Production never matches the development environment. It resembles, but cannot match

So they deployed people on the other side

3. But this process was slow, they wanted to deploy faster

So they deployed Continuous-Deployment (CI/CD)

4. To improve reliability, we got SRE to do this

SREs’ first job was to hold this ship, but that’s all where they got stuck at

5. What Site Reliability Engineers should’ve built is

SREs should’ve been *engineering* and *observing* the bridge, but instead they became the bridge

Share to:
Twitter
Reddit
Linkedin
#SRE Tooling #sre #devops #Observability #SLO #Deep Dives #Last9 Engineering #Last9 #Failures #hans #tools #Systems Engineering #Latency

You might also like...

Latency SLO
Latency SLO

How do you set Latency based alerts? The most common measurement is a percentile-based expression like: 95% of the requests must complete within 350ms. But is it as simple?

Read ->
A Primer on Saturation SLO: What Is It and Do You Need to Consider It?
A Primer on Saturation SLO: What Is It and Do You Need to Consider It?

What is Saturation and why should you think about it as an SLO? Saturation can be understood as the load on your network and server resources.

Read ->
Sleep Friendly Alerting
Sleep Friendly Alerting

We've all been woken up with that dreaded Slack notification at ungodly hours only to realise that the alert was all smoke and no fire. The perfect recipe for dread and alert fatigue.

Read ->

SRE with Last9 is incredibly easy. But don’t just take our word for it.