Where Platform Engineering GOES WRONG

Gregor Hohpe talks to Dave about Platform Engineering. They discuss where platforms go wrong and the common mistakes organisations make when using it to guide software development.

___________________________________________

🙏The Engineering Room series is SPONSORED BY EQUAL EXPERTS

___________________________________________

📚 BOOKS:

LINKS:

#podcast #softwareengineer #platformengineering

[zoltannemeth]

Our approach is to set defaults that reflect the governance (policies, cyber security compliance) as this *does* help developers immensely compared with “trial and error” approach to see what is allowed and what is not. Where governance allows flexibility, the defaults can be overridden by the developer, but they are making a conscious choice to do this and therefore are responsible for the consequences of their actions.

For example, creating “globally redundant storage” is not our default setting for our templates; locally redundant storage is. The developer / architect can certainly override that setting, but they are expected to defend the resulting increased costs to the business. (I.e. if this provides DR coverage for a solution’s important data, this decision is completely defensible, however, doing this for a non-prod test instance of the app, most likely is not defensible.

[gardnerjens]

to me if the platform can take remove the need for a ticket system, through a developer oriented experience a lot is achieved,

I agree, to me when i see people working on a platform, it more starts to look like a rod to control what the developer teams are doing so they dont go out of lockstep.

Where as i wish the platform be a means of tools to ease the process for the developer teams.

[adambickford]

Almost everything that a 'platform' brings is taken away by the inability to google it, ramp up new users, etc. even if it's good.

I've never seen a good one irl.

[tafkamextra]

Please make an episode on how ticketing systems are used in companies the wrong way stifling progress and success!

[yannick]

It all comes down to having the right building blocks/primitives. So far most platforms I had the pleasure to work with were really disappointing: complicated, slow, inflexible and expensive. Lack of support in case of problems or questions as well. Kind of the same problems we had with ticketing systems. But what benefits to I get except for getting it nearly instantly on the click of a button? That is such a small time frame for several month or year long projects. Additionally I now no longer have a human on the other end that can help. If the process is bad it doesn't help to automate it. In many businesses the process has not changed.

[evolagenda]

I fundamentally disagree. This assumes an idealised scenario. There's a huge spectrum of maturity in organisations. Any where that's not a elite shop absolutely needs to be provided an abstract route to live.

[kaes]

I came here only to say that THERE IS NO SHORTAGE of software engineers

[lost-prototype]

Ops is dead in a lot of ways. Maybe not in spirit, but in role.