What about Telemetry, something that is quite specifiic to BEAM ecosystem?
kokolegorille
cnck1387
Most popular languages / frameworks have their own features and libraries to get similar stats.
For example with Flask, Rails, Laravel and Django there’s various forms of “debug toolbars” that give you stats like how long a request took, how many DB queries you had, how long those queries took, how much time was spent compiling a view, queries against Redis, etc..
These are drop in solutions requiring zero lines of code and are extensible to add your own custom stats. There’s also language specific tools for profiling function calls and getting break downs of how long it took to execute a specific part of a function, etc..
Then there’s also a bunch of APM tools out there for nearly every language that’ll let you get in depth stats about pretty much anything you could imagine. All results are persisted too, so you can compare and chart the results over time.
System specs like memory / CPU usage, etc. is also a solved problem with many dozens of tools. Often times you can get a moving chart (with weeks of persistence) of this straight from your cloud provider for free and without having to set anything up. DigitalOcean will do this for you.
1
kokolegorille
cnck1387
Yeah totally, like other languages Elixir has its take on metric gathering too.
Originally you mentioned the idea of telemetry is unique to the BEAM. I was just trying to say that lots of other languages have the ability to get various metrics out of their runtime and then display those metrics in easy to digest ways. It’s just not called telemetry in most other languages.
1
kokolegorille
I speak about telemetry, and You speak of other frameworks… I don’t see the relation.
Have You ever implemented telemetry in Elixir?
It is specific to the BEAM how it is done.
0
olafura
Hi Isaac,
thank you for coming here and being curious about our discussion of the project.
I haven’t written code for the debian Benchmarks Game yet. But yeah. I was looking at GMP. Not realizing that Erlang uses GMP under the hood 
and it also uses PCRE.
So it would probably be interesting seeing if we can bring the speed closer to the top.
It’s really hard making these kinds of benchmarks because each language has their own way of doing things that might not align with those kinds of benchmarks.
So I’ll probably try to contribute Erlang or Elixir examples if I can find some speed ups instead of complaining about them on a forum 
0