I think a fair summary is "We had a few problems with PostgreSQL mostly due to our rapid growth. We rethought the problem and changed the way we use a relational database for large-scale storage and are now using MySQL as a dumb key-value store."

The conclusion reinforces this: "Postgres served us well in the early days of Uber, but we ran into significant problems scaling Postgres with our growth."

I read this a both endorsement of PostgreSQL as well as highlighting some of the problems that any large-scale use of it would run into.

June (https://juneoven.com/) is hiring! | Full-time | SF

June is a team on a mission to transform the kitchen. Our first product is the June Intelligent Oven. A built-in camera and sensors makes June able to identify commonly cooked foods, recommend a cooking program and execute to perfection. With a touch screen, instant-on carbon fiber heating elements, advanced moulded insulation and a finely tuned cooking cavity, we’re bringing restaurant-level performance to the home.

We're looking for Software and Hardware engineers.

See https://juneoven.com/jobs.html

June is hiring! — San Francisco, CA

We just came out of stealth and are looking for smart and creative software and hardware engineers to join us on our mission to bring innovation to the kitchen and home.

Specifically, we're looking for:

  • Android UI Engineers
  • Embedded Engineers
  • Mechanical Product Design Engineers
  • Hardware Engineers
  • Server Engineers

Send us your resume at https://juneoven.com/jobs.html

★★★ June is hiring! — San Francisco, CA ★★★

We just came out of stealth and are looking for smart and creative software and hardware engineers to join us on our mission to bring innovation to the kitchen and home. Specifically, we're looking for:

  • Android UI Engineer
  • Embedded Engineer
  • Mechanical Product Design Engineer 
  • Hardware Engineer
  • Server Engineer

Send us your resume at https://juneoven.com/jobs.html

Agreed. I think this title is a bit flippant and I'm sure the 600M+ WhatsApp users don't care what the servers are running.

The important takeway is that FreeBSD and Erlang have worked really well for WhatsApp allowing them to grow their userbase insanely fast while still staying relatively small.

This should no longer be true. Chrome v31 (http://googlechromereleases.blogspot.com/2013/11/stable-chan...) added SCTP support for reliable DataChannels and allowing compatibility with Firefox.

Chrome and Firefox DataChannels cannot connect with each other. Chrome and Firefox media connections can interop though

This is a really nice writeup of the whole startup process. Thanks for this!

Bottle uses `wsgiref` (https://github.com/defnull/bottle/blob/master/bottle.py#L263...) by default as its server. Per the name, it is a reference implementation of a WSGI server. No one should really use `wsgiref` for an application they care about. This test is invalid for this and lots of other reasons.

I suspected many in the community here would host their own XMPP server.

What XMPP server are you using or would you use?

I've had great luck with prosody. I used to use jabberd2, but I switched away from it due to its dependencies on various badly-maintained libraries.

We also use Prosody and like it. For us it's been very close to an "it just works" solution.

The SQLite site continues to offer a wealth of information about SQLite internals. Most of the pages (ex. http://www.sqlite.org/fileio.html) are a good read for anyone interested databases or creating good system software.