texascloud

I really like the points about round-robining team members, and timeboxing confusion. I think the post does a good job diving into some of the trickier parts of working in a development team besides "can I understand the code"

BTW, I highly recommend reaching out to the solo.io folks cuz they can help unravel the docs very quickly and get you pointed to where you need to go! Digging in is worth it too!

My tip: look at the .proto files on the docs site. They map (I think) 1:1 with the CRs, and explains what each field is for

Thanks!

I think Ambassador has better self-service CRs that function together well without needing to define explicit dependencies. An example is TLSContexts and Mappings. As long as you have a host value in your mapping which matches exactly the host value of a TLSContext, that cert is used. With Gloo, it's structured differently altogether, but they have no separate resource for tls/ssl. It's defined in the VirtualService itself, so strongly coupled to the routes. Route delegation (in newer gloo) helps with this a lot!

I believe the gloo GitHub has an open issue for making ssl config its own CR, which should help.

One other thing is Ambassador (starting in 1.0) allows Mappings to be created that refer to Kubernetes Services by name, using the Namespace the Mapping is created in to find that Service.

This is incredibly helpful when you have the same services but in different Namespaces (like with multi-tenant clusters), because the same set of Mapping YAML can be applied in each Namespace without having to first generate them from templates. It depends on how you deploy.

If you're not planning to have a huge cluster with lots of dynamically added routes, and don't already have complex deploy logic, I say it's fine to keep using Ambassador. Although due to the complexity of switching ingress controllers, standardizing on one is worth it IMO.

About the Edge stack, I did try to use it, but I was unable to successfully get it to deploy in a fresh cluster. This was literally the week of them launching it, so stuff was chaotic. I also had modifications to the codebase that were required for me to use it effectively, and Edge stack is a special container that has things baked into it, so I couldn't simply change the Deployment YAML to use my container image based on the OSS version :(

This wasn't a strict controlled test with perfect setup. It was a personal account of a busy developer who works at a start-up trying to kick the tires on two technologies in a realistic way in order to decide which to move forward with as we faced scaling challenges.

I hope HN appreciates the scrappiness more than they dog on the sloppiness :)

What is MLM?

Agreed. It's dangerous. This is simply someone's personal experiment and their experiences written up. It's proof of nothing other than "I did something"

It's so strange to see "independent researchers" attempt to poke holes by any means in an argument meant to protect people from the ill effects of sleep deprivation, yet provide zero non-anecdotal evidence to support lack of sleep being beneficial.

Discounting Walker's argument with "he's not thorough" and then providing a "study" done by one person with no professional research experience appears fine for HN

Proof of digging?

Okay, he removed the 5 hour bar. I get what the author in your link is upset about, leaving out data points. But imo that's not good enough to make me throw out all the information in that book.

It sounds like the author is upset a book about getting enough sleep is getting so much attention. If they're upset at the way Walker misrepresents information, then I'd hate to see the author's blood pressure while feverishly researching the bias of every article they come across

I wouldn't say this is refuting even a minority of claims in the book. It appears to be a thorough fact check of a small percentage of claims, some of which is going after the author over wording and inconsistencies. It's fair criticism, but by no means invalidates a majority of the information regarding sleep that Walker presents

This 100%. I've read the book twice, coming back to certain chapters more than twice, because it's dense with information. It's scary really. The single most impactful thing you can do as a human for your life is to prioritize 8 hours of sleep per night, full stop.

Are there cameras on the sides of the occulus so peripheral hand movements can be recognized?

Only reading the abstract, are you saying a different branch/fork of Linux could be maintained to be a unikernel?

I'm personally really excited for this! I wonder just how much slower complication is versus Go. Would it be enough to matter with a large project?

Is this good for IBM?