Sandro Manke

Architecture, Dev, Ops, Engineering, Platforms, Systems, Public/Hybrid Cloud, Security, Governance, etc.

Freelance Software & System Architect / Engineer

MHP – A Porsche Company

Freelance Architect / Engineer

50Hertz Transmission GmbH

Smartbroker AG

Asklepios

Better Payment Germany GmbH

Swissbit AG

Qodeup

Hermes Europe GmbH

Porsche AG

Turning ☀️ into ☁️ since 1999.
Technology is a means to an end to achieve business goals. Nothing more. 
I believe that tech does not solve problems, but we do and tech can support us. 
We as humans are responsible for finding simple - not easy - solutions with the least tech necessary.

Kubernetes: Wann ist es wirklich notwendig? 🤔

Ich erinnere mich noch gut daran, wie fasziniert ich war, als ich das erste Mal von Kubernetes hörte. Die Idee, dass der Betrieb und die Operations von Anwendungen über mehrere Firmen hinweg ähnlich werden könnten, klang verlockend. 🌟

Doch in den letzten Jahren habe ich viele Teams gesehen, die einen Kubernetes Cluster betreiben und damit eigentlich komplett überfordert waren. Der Cluster war oft viele Versionen hinter der aktuellsten zurück und keiner hat sich getraut, das Update zu verantworten.

Und dann kam es, wie es kommen musste: Der DevOps Architekt, der alles gebaut hatte, verließ die Firma und das Team stand dumm da. 🙈

Aber brauchen wir Kubernetes wirklich in den meisten Fällen? 🧐

Kubernetes ist ein tolles Stück Technologie, keine Frage. Aber am Ende des Tages sind die meisten Anwendungen nicht groß genug, um den Einsatz von Kubernetes zu rechtfertigen. 

Der Betrieb eines Kubernetes Clusters ist extrem aufwändig und erfordert dediziertes Personal. Selbst die managed Solutions von Cloud-Anbietern wie AWS EKS sind nicht einfach "Fire and Forget". 

Gerade Cluster Updates, die häufig kommen, sind kein Spaß. Und auch das Argument, man könne einfach von einem Cloud-Anbieter zum anderen wechseln, halte ich nicht für so valide. 🚫

In vielen Fällen können die etwas simpleren managed Container Plattformen Sinn machen. Bei AWS bevorzuge ich oft ECS Fargate, weil ich mich nach dem Deployment um nichts mehr kümmern muss. 😌

Fazit: Kubernetes ist toll, aber oft Overkill. Überlegt genau, ob ihr es wirklich braucht! 💡

Was denkt ihr? Habt ihr ähnliche Erfahrungen gemacht?

Kubernetes now offers microtransactions. 
Your pod crashed? 
Pay up! 

There is no 10x, but beware of the unprofessional 0.X externals!

This does not apply to everyone, and I see this barely with employees. There are also companies where you wouldn't find this at all. But it's still so widespread that it is simply scary.

I primarily see this with labor leasing (all those "agencies") and some "freelance" contractors (mainly as freelance does not mean the same for everyone)

"Barely Juniors" are being sold for a lot of money who think they are a lot better than they actually are, with zero skin in the game and no drive to actually deliver as they get paid even if they do not deliver anything at all. 

It would be okay if they delivered slowly. 

The speed does not matter as long as it moves things forward.

You know, >1X.

But what happens is that they actively slow down everyone around them with complaints, why they can't work, 💩ing in the codebase, which everyone else has to first debug and then clean up, again complaining how things need to be refactored and changed and how the culture sucks, without doing anything productive. In contrast, others are productive around them.

Hire enough of those, and eventually, no one will be productive anymore. 
Seen this more than I can count.

I agree that culture matters. Many things matter. 

You are brought in to move things forward in the mess it is (if it weren't a mess, you'd not need temps) while iteratively improving it.

You are *not* supposed to be spending all that time rewriting it how you think it's better for everyone in isolation while nothing at all moves forward and the outcome is just the next questionable 💩  that doesn't work. 

Improving things is part of normal work. If that means stopping to work for weeks or months, then sorry, you are doing it wrong. 

You can move forward and improve things, always, at all times. 

That may make it harder, but in a business, it is critical to ensure that whatever you build is tested early and does not interrupt other people's jobs. The business depends on that.

I wish there was more professionalism in the business. 

I am aware it's not a single person's fault, and it's a complex scenario where this is thriving, but it hurts everyone unnecessarily and eventually makes externals seem bad. Companies and hiring policies play an important part here, but that does not make offering horrible "services" like this any less questionable.

This dumb game hurts everyone involved.

Stop that.

Taplio

Sandro Manke's Linkedin Analytics

Sandro Manke

Sandro Manke's Best Posts (last 30 days)

Want to drive more opportunities from LinkedIn?

Taplio

.css-1jgreyo{background-image:linear-gradient(284deg, rgba(39,43,72,1) 0%, rgba(33,47,183,1) 100%);color:transparent;-webkit-background-clip:text;background-clip:text;}Sandro Manke's Linkedin Analytics

Sandro Manke

Sandro Manke's Best Posts .css-qqfgvy{font-size:var(--chakra-fontSizes-md);}(last 30 days)

Want to drive more opportunities from LinkedIn?

Sandro Manke's Linkedin Analytics

Sandro Manke's Best Posts (last 30 days)