swaranga

Lots to agree; I am certainly feeling it these days. I used to be excited for new language features (records, pattern matching) but I was looking at the recently posted Devoxx conference talks on Youtube and for every one of these, I was thinking "what's the point? The code will be written by an agent anyway".

Some Linux?

I keep forgetting the reasoning for this part of the design - why is the value semantics baked in the declaration site and not at the use site?

Why does the class author have to declare it as a value type? I know this has been rehashed before many times by Brian but I keep forgetting the reasoning.

Amazon may be big but even then resources are limited. Individual teams owning the roadmap for services are still small. And there are always competing priorities. Do you prioritize these security and availability improvements or do you pickup some operational improvements to make our engineers oncall easier? Oh there is also this long line of customer features you need to deliver. And re:Invent is also approaching.

When you look at it like this, not very surprising that initiatives like cost savings optimizations may take a back seat for periods of time.

For me it is the little things. Like switching windows. Mac feels half a second slower then Linux and that delay make it feel less responsive. Opening up the terminal can iterm2 is another one. For me these things just adds up. I want my OS to get out of my way when I want to get work done.

I have turned off the genie effect and perhaps other animations, but my Ubuntu on my Lenovo still feels faster than the Mac.

Sure, it works for a narrow use-case like a pure function that just queries and returns data. What about stored procedures/triggers that mutate data on insert/delete etc?

With one-box, you can assign a specific set of customers/canaries to be served by the new 1-box and validate your results. If there is a problem you rollback and only your canary data was affected. It is just not very simple IMO.

Not a security engineer so can’t comment on the threat model. But an insider threat like a rogue employee can do as much damage with a managed laptop vs an unmanaged one IMO.

I am loving the Netflix model. You can ask for a MacBook or a thinkpad. If you opt for the thinkpad, you can also install whatever Linux you want and all tools will still work.

In fact you can bring your own laptop and without reimaging use that for development. All interval tools (of which there aren’t too many are written in Go so are cross platform). I ordered a framework laptop and will expense it once it is here.

Onebox deployment typically deploys the new revision of the application code to a single host/box and leaves the rest of the fleet in the older revision and then you run your tests against the new revision running in that single box. You also might monitor the metrics from that single box and attempt to find regressions from the rest of the fleet.

If the code is deployed to the database (version controlled or not) all application servers will start using the new code immediately which is not what we would want from a onebox deployment first strategy. Usually all the servers will share the same database and thus the code in the code in database model. What am I missing.

How would you do one-box deployments with code in the database? Everything else could be solved with perhaps some or a lot of pain.