Tweet

Steve Smith

1 Jun, 11 tweets, 9 min read

Posted: GitOps is a placebo stevesmith.tech/blog/gitops-is…

Why #GitOps:

1. is just a rebadging of #ContinuousDelivery and #InfraAsCode ideas
2. has no new ideas of substance
3. contains 'best practices' that won't always work
3. offers no benefits that can't be achieved without GitOps

https://twitter.com/SteveSmith_Tech/status/988813078128099329

A few words on GitOps is a placebo stevesmith.tech/blog/gitops-is…...

I've disliked #GitOps from the start

https://twitter.com/SteveSmith_Tech/status/988813078128099329

. It went from "startup sharing ideas that worked for them" (good) into cargo culting (bad)

Git, Kubernetes, <insert tool here> are not special

#ContinuousDelivery and #InfraAsCode are tool agnostic, they're about principles and practices. They matter.

When someone says they've combined those principles and practices with some tools, to "extend" those principles and practices... I become cross.

Git will be replaced (we don't all need a swiss army knife for time machines)

Kubernetes will be replaced (we don't all need to pilot our own giant container ship with a giant learning curve)

#ContinuousDelivery was here before Git and Kubernetes, it'll be here after

https://twitter.com/SteveSmith_Tech/status/1083149478679465984

I don't normally write articles that just say "X is bad", I try to say "X is bad, Y is better, here's how to move from X to Y"

I delayed writing about #GitOps for 3 years because I couldn't figure out Y.

https://twitter.com/SteveSmith_Tech/status/1083149478679465984

@meekrosoft

My thanks to @meekrosoft for suggesting 'placebo' as a more accurate term for #GitOps than 'snake oil'. It's a sugar pill not snake oil, it has benefits, it's not fraudulently peddled, the problem is its benefits are wholly transitive

https://twitter.com/SteveSmith_Tech/status/1094160494091485184

#GitOps isn't needed.

You can achieve the same benefits with Subversion, some Python scripts, and Nomad. You might choose that, but it's fine. They're just tools

And GitOps can lead to branching problems incompatible with #ContinuousDelivery (like Git, funny that)

@adamhjk

Thanks to @adamhjk, @ringods for occasionally popping up with "is anti #GitOps a thing?"

https://twitter.com/SteveSmith_Tech/status/1243113319604588544

Yes it is, I'm driving, all aboard, plenty of room

@sigje

Also thanks to @sigje @IanColdwater for their parallel anti #GitOps train, we can converge now

https://twitter.com/SteveSmith_Tech/status/1392504433460891648

https://twitter.com/SteveSmith_Tech/status/1396572548868067335

If #GitOps was renamed as 'Infrastructure as Code for Kubernetes'

https://twitter.com/SteveSmith_Tech/status/1396572548868067335

and

Weaveworks stopped their plans for #GitOps certification

https://twitter.com/SteveSmith_Tech/status/1397350218199408640

The world would be a little bit better

@davefarley77

Thanks to @davefarley77 @KrisBuytaert @tdpauw for reviewing GitOps is a Placebo stevesmith.tech/blog/gitops-is…

It involved a lot of late nights, re-reading CD/IaC books and GitOps materials. No current work colleagues were involved, my opinions are my own, #GitOps sucks

• • •

Missing some Tweet in this thread? You can try to force a refresh

This Thread may be Removed Anytime!

Twitter may remove this content at anytime! Save it as PDF for later use!

More from @SteveSmith_Tech

Steve Smith

@SteveSmith_Tech

30 Jul 20

How should you decide if an incident merits a post-incident review?

The answer isn't "if it's a P1 or P2, forget P3s". In fact, it's the wrong question to be asking... #Operability 1/n

Many orgs I work with have a policy of "do a review if the incident was a P1 or a P2". Lower priority incidents don't get a review.

This might be due to a high volume of P3s from untuned alerts, friction in the incident review process, lack of emphasis on improvement, etc. 2/n

@AdaptiveCLabs

At best, the incident review process involves team members working together to uncover a shared timeline and improvement actions.

I'd call this a "shallow analysis" that pre-dates an understanding of resilience engineering, operability, the work of @AdaptiveCLabs etc. 3/n

Read 13 tweets

Steve Smith

@SteveSmith_Tech

25 Jun 20

@clare_liguori

A deep dive into @clare_liguori's great article on #ContinuousDelivery at @awscloud... 1/n

https://twitter.com/clare_liguori/status/1273695334066057216

@clare_liguori

Firstly, thanks to @clare_liguori for doing this. As someone with deep knowledge of deployment pipeline design and shallow knowledge of AWS, this is really interesting. I wish more companies did this 🙇‍♂️ 2/n

As someone who's been thinking about deployment pipeline designs for over a decade e.g. continuousdeliveryconsulting.com/blog/deploymen… and vimeo.com/370035221, it's always interesting to hear how orgs do deployment pipelines a) at scale and b) in unusual market conditions. AWS is a) and b) 3/n

Read 31 tweets

Steve Smith

@SteveSmith_Tech

21 Aug 19

Thread!

If you've decoupled deployments from customer launches, well done 👍

*But* bear in mind #ContinuousDelivery is about increasing throughout to meet customer demand. So it's the frequency of customer launches you need to increase, not just deployments 1/n

Deploying more frequently to production reduces technology risk, overheads, inefficiencies, etc. It's a good thing.

However, it's customer launches that create opportunities for validated learning 2/n

#ContinuousDelivery was named after the first principle of the Agile Manifesto agilemanifesto.org/principles.html

"Our highest priority is to satisfy the customer through early and continuous delivery of valuable software"

It is a means to achieve customer satisfaction 3/n

Read 7 tweets