HN user

jayp

457 karma

Previously Co-Founder of Herald and also Co-Founder of Midtype (YC S19)

Posts10
Comments127
View on HN
Monorepo Support 4 years ago

If you are going to use a centralize database coupled with edge compute, you may end up with far worse end-to-end latencies if your compute<->db does multiple roundtrips per request (which is fairly common in practice). In most cases, you need compute and db to be colocated.

I am skeptical of edge compute being generalizable. I do forsee a bright future for it for embarrassingly parallel problem space, where data sharding can be done cleanly based on an end user.

They and Vercel both seem to be going thru a lot of growing pains. Quite a few outages over the past year.

PS: we use both (for different sites). Probably should consolidate to one.

Hi Kevin ;-)

I am founder of another YC backed company. We based our startup on GCP infra. They have great tech (for the most part) but I regret it so deeply for two reasons:

Support or desire to help customers is non-existent. For any questions, they want us to upgrade to paid support and pay them at least 10% more every month for that (we ask like 1-2 questions a year). What? We are already paying you thousands of dollars every month! I get included support for all my software subscriptions - so this is my biggest beef. Also, when they do help, they keep passing you around from team to team and dont resolve issues as well I’d like. It is just not a company I can love as a customer.

Their status dashboard is a joke. They dont even report minor outages, when they do, they start after a huge delay and update very slowly. And worst of all - when it only affects a single zone or a single region, they remove it from historic reports so everything looks green/great.

I’ve experienced both these times multiple times.

I have to assume AWS is better.

Good question. Probably the same. English, and especially business English, is not a precise language. I am not really an expert but was part of the larger org.

Just to be clear - colocation here is in order of thousands of square feet. The datacenter provider provides redundant utilities. The customer does everything else.

Where is this place? I agree with grandparent - for the past decade or so, having your “own datacenter” means leasing floorspace in an actual datacenter run and managed by someone else.

I live and work in the Bay area. I was part of the larger org that ran Uber’s datacenter (co-located) and that’s how it was talked about.

Netlify announces move from .netlify.com to .netlify.app with the note that "it’s important to note all of your existing sites will continue to operate properly using their current URLs." Even says so here now: https://community.netlify.com/t/changes-coming-to-netlify-si....

What is their solution? Just fwd .netlify.com to .netlify.app. This breaks Chrome extensions that hotload code from a specific domain. The Fix? To publish a new version to Chrome webstore to unbreak this. The approval takes a few days.

Netlify obviously did not think of all corner cases or allow users to test this before hand!

Thought I'd write up a technique that's helping iterate a lot faster on our startup. Would be happy to answer any questions or entertain thoughts I may have missed.

This comment is aligned very much with my current thinking. Hiring highly paid engineers but not maximizing their problem solving ability is a significant problem in large companies.

Empowering bottoms-up problem solving by sharing customer pains and feedback broadly is the exactly problem I am trying to solve with my current project/startup [1].

I am gonna shoot you an email later today to see if you have any further thoughts on how do this effectively within an org.

If anyone has ideas on good implementing this, would love to learn so we can incorporate this within our offering. My email is in my profile.

[1] https://www.heraldhq.com

Uber has always been self hosted. Some workloads are on Cloud and migrating more there. I last worked there 2 years ago.

My co-founder at Midtype, Neel, wrote this article by looking at onboarding flows of a dozen or so successful companies.

We have picked up a few tips on how to improve our own onboarding process.

What is a good onboarding that you experienced recently?