AWS Lambda PR/FAQ After 10 Years | Hacker News Reader

AWS Lambda PR/FAQ After 10 Years

allthingsdistributed.com

27 pointsmlhpdx12 comments

Let's see... Here's how I imagine my old management reacting to this PR/FAQ:

- The press release needs to be more customer focused. Can you justify "simplest"? Please add details about which specific AWS services can work with Lambda. Why are you mentioning Python and Ruby? Please remove 250ms.

- The customer quote should name a real person, not XXX. Where did the 1M number come from? Please add details to the Appendix. Cut part about "huge management and operational overhead". Please use specifics instead of "fast" and "reliable".

- FAQ #1: What is "stateless"? Please expand. Why is this different from other AWS services? Provide real numbers and examples.

- FAQ #2: This is unclear. What about other potential Lambda users? Provide real numbers and examples.

- FAQ #3: Everything below "handlers" is too specific, please merge with #4.

- "Developing and Deploying" section should be moved down. Security section should be first, as it's the priority.

- Overall, too much technical detail. Clean up and add to Appendix. Focus FAQ on customer needs, not implementation details.

- Etc.

Note that all of these comments will be contradicted by another manager in the following week's overview meeting.

All valid points. And having been at Amazon for two stints, I think it's fair to say that it really depends on who's reviewing the PR/FAQ.

I did review this PR before Werner published it, and can say that in this specific case, attributing the quotes to anyone for public release was not allowed. However, I would definitely expect the name of a real person for any quote in a PR/FAQ I was reviewing internally.

Those aren't valid points!!! That wasn't my message at all. They're idiotic nitpicking points from non-technical managers whose only job is to pedantically edit PR/FAQs as if they mean anything. And each manager has their own opinion about what should or should not be in one.

That document is perfectly fine. It clearly expresses both the idea and the business value of it, which has obviously been borne out by the unmitigated success of Lambda.

The whole company has become brainwashed by moronic MBAs. The amount of stupidity I dealt with writing and re-writing PR/FAQs has taken years off my life.

The only real message my comment has is to show the poor suckers still working at Amazon the bullshit that they should expect when writing one. Thank all that is holy I never have to deal with those idiots ever again.

Some of these points are valid though. Because you should to some degree challenge the author to simplify where possible.

Does it get out of hand. Absolutely. But calling out when a PR/FAQ is too technical is valid. Some details belong in a design doc (or elsewhere).

However, I agree with you that the intent should be to deliver a good product. Not to nitpick someone into an anxiety spiral.

Glaring grammatical error in the PR headline? :)

Amazon Web Services Launches AWS Lambda

Easily run any code as an AWS-operated scalable, secure, and reliable cloud services

Is Lambda really relevant? Ten years later I think most businesses are skeptical of becoming dependent on a proprietary service as opposed to general compute resources.

It is a powerful capability for executing arbitrary user provided code in a secure manner. You can then repatriate to OpenFaaS on Kubernetes if an exit strategy is required.

I'd guess the answer is that "yes" it is still relevant, but I'd be guessing that "most" companies care more about economic efficiency than ideology, too. If you have metrics on how much companies spend to be "skeptical" I'm all ears.

I’m curious how they knew EC2 instances were sitting idle waiting to process a file.

I would guess host metrics show they're idle, and talking to say 10 such customers showed obvious patterns. Is it harder than that?

AWS SA here, but speculating, so take it with a grain of salt.

A lot of customers -- especially the larger ones -- actually share their architecture and have in depth conversations about what they are doing and pain points, etc, with their account teams.

My guess is that they knew because their customers told them (or complained?) that they had fleets of EC2 instances just waiting

Wouldn't AWS have metrics about physical and virtual CPU utilization from the hypervisor?

Reply on news.ycombinator.com