5 ms·

I've worked at a couple places that had developers build their own environments before I got there (I'm more of a systems engineer). There was no thought to hi

I've worked at a couple places that had developers build their own environments before I got there (I'm more of a systems engineer). There was no thought to high availability, security, scalability, reliable backups, etc. I'm not saying all devs are like this, but that's just been my experience.

Thaxll 8y ago
This. Honestly Ops is a different mindset that devs, the tldr version is devs don't really know / care about running things in production. I don't blame them it's just different set of skills.
- dtwwtd 8y ago
  > devs don't really know / care about running things in production I disagree with this entirely. Devs _you have workd with_ might not have cared but it doesn't make the statement universally true. As a counter argument, some of the best software developers I've worked with were also very good at operating and debugging production software and the reverse has also been true. I also don't buy that the two are a different mindset (at least in the domains I work in and care about). In my experience the very best people (whether they're working "dev" or "ops" roles or something in the middle care about the entire development and deployment lifecycle of the software they work on. Building a good experience in software includes thinking about reliability and availability and planning a reliable and performant deployment also requires that to be thought of in the application layer at some point.
- jteppinette 8y ago
  A developer who does not know/care about running things in production is not a developer. These are not different skills. Security, high availability, reliability, these are concepts that hold as much weight at the application side as they do on the systems side. Two sides to the same coin.
  - saint_abroad 8y ago
    > A developer who does not know/care about running things in production is not a developer. Behaviour depends on how incentives and constraints are structured. Traditionally, operations had to keep things running and cared little about new functionality. Developers had to deliver new functionality and cared little about keeping things running. Requiring developers attend 3am outage calls changes incentives, as does requiring operations to own (or get out of the way of) software delivery pipelines.

Thaxll 8y ago

This. Honestly Ops is a different mindset that devs, the tldr version is devs don't really know / care about running things in production. I don't blame them it's just different set of skills.

dtwwtd 8y ago
> devs don't really know / care about running things in production I disagree with this entirely. Devs _you have workd with_ might not have cared but it doesn't make the statement universally true. As a counter argument, some of the best software developers I've worked with were also very good at operating and debugging production software and the reverse has also been true. I also don't buy that the two are a different mindset (at least in the domains I work in and care about). In my experience the very best people (whether they're working "dev" or "ops" roles or something in the middle care about the entire development and deployment lifecycle of the software they work on. Building a good experience in software includes thinking about reliability and availability and planning a reliable and performant deployment also requires that to be thought of in the application layer at some point.
jteppinette 8y ago
A developer who does not know/care about running things in production is not a developer. These are not different skills. Security, high availability, reliability, these are concepts that hold as much weight at the application side as they do on the systems side. Two sides to the same coin.
- saint_abroad 8y ago
  > A developer who does not know/care about running things in production is not a developer. Behaviour depends on how incentives and constraints are structured. Traditionally, operations had to keep things running and cared little about new functionality. Developers had to deliver new functionality and cared little about keeping things running. Requiring developers attend 3am outage calls changes incentives, as does requiring operations to own (or get out of the way of) software delivery pipelines.