3 ms·

A series of fatal errors caused the majority of servers to become unavailable. We are working to revive all of these resources. Most users will be unable to con

A series of fatal errors caused the majority of servers to become unavailable. We are working to revive all of these resources. Most users will be unable to connect while this work is ongoing. That's microservices for you. Instead of one fatal error you have to fix there's multiple in different codebases at the same time

kyleee 7y ago
Can you remove the code formatting on that and just use quotations? Very difficult to read on mobile
- caleb-allen 7y ago
  Here is what it says: "A series of fatal errors caused the majority of servers to become unavailable. We are working to revive all of these resources. Most users will be unable to connect while this work is ongoing."
raziel2p 7y ago
That's jumping to conclusions very quickly. I could easily say the same thing about a monolithic spaghetti codebase where a change in one place led to unintended consequences somewhere else.
conradfr 7y ago
Are they using microservices though? I think it's Elixir+Rust nifs and go/c++ for media. And we have zero informations on what the problem is.
- hinkley 7y ago
  "Any sufficiently complicated concurrent program in another language contains an ad hoc informally-specified bug-ridden slow implementation of half of Erlang." Akka, Microservices, half of Kubernetes... all are making similar observations and coming to different conclusions and/or battle plans based on the tools near at hand. But it's not like monoliths don't have huge, even bigger problems with action-at-a-distance. The problems that get you are a mix of simple human errors and subtle patterns of side effects that are difficult to spot or predict. Simple errors are easy to fix wherever. Emergent behavior is much harder to cope with.
  - t-writescode 7y ago
    Elixir is a language on top of Erlang, if memory serves me, so yes, Erlang is in there :)
    - hinkley 7y ago
      Yeah, it is, and it’s at the top of my todo list.
1123581321 7y ago
Your username made me laugh. :) I haven’t found multiple service architecture to have this issue in a careful, sufficiently large engineering org unless they share vulnerable infrastructure. There should not be any situation where multiple APIs are changing in a backwards-incompatible way simultaneously.

kyleee 7y ago

Can you remove the code formatting on that and just use quotations? Very difficult to read on mobile

caleb-allen 7y ago
Here is what it says: "A series of fatal errors caused the majority of servers to become unavailable. We are working to revive all of these resources. Most users will be unable to connect while this work is ongoing."

raziel2p 7y ago

That's jumping to conclusions very quickly. I could easily say the same thing about a monolithic spaghetti codebase where a change in one place led to unintended consequences somewhere else.

conradfr 7y ago

Are they using microservices though? I think it's Elixir+Rust nifs and go/c++ for media. And we have zero informations on what the problem is.

hinkley 7y ago
"Any sufficiently complicated concurrent program in another language contains an ad hoc informally-specified bug-ridden slow implementation of half of Erlang." Akka, Microservices, half of Kubernetes... all are making similar observations and coming to different conclusions and/or battle plans based on the tools near at hand. But it's not like monoliths don't have huge, even bigger problems with action-at-a-distance. The problems that get you are a mix of simple human errors and subtle patterns of side effects that are difficult to spot or predict. Simple errors are easy to fix wherever. Emergent behavior is much harder to cope with.
- t-writescode 7y ago
  Elixir is a language on top of Erlang, if memory serves me, so yes, Erlang is in there :)
  - hinkley 7y ago
    Yeah, it is, and it’s at the top of my todo list.

1123581321 7y ago

Your username made me laugh. :) I haven’t found multiple service architecture to have this issue in a careful, sufficiently large engineering org unless they share vulnerable infrastructure. There should not be any situation where multiple APIs are changing in a backwards-incompatible way simultaneously.