Configration and general questions about nameko HOT 2 CLOSED

Hi,

1. I believe the reasons for a relatively low default were
   a) the initial use-case (onefinestay) didn't have a huge number of concurrent activity, and
   b) while the work was mostly io-bound, there were tasks which used a non-trivial amount of cpu, which "starves" all other greenthreads. This was mitigated by not having too many concurrent workers. Depending on your use-case, a much larger number of concurrent workers may work fine.
2. Tested in the sense that it's been run in production environments that deal with sizeable amounts of data without any noticeable increase in long-term memory usage over time. During its development there have been memory leaks, but to my knowledge they have all been fixed. Of course no software is perfect but a memory leak would certainly be considered a bug.
3. Nameko has been used in production at onefinestay for several years at this point, running the main backend systems.
4. I generally run it with supervisord. If you launch your services with nameko run myservice, the runner installs a signalhandler for SIGTERM that requests a graceful shutdown (stops accepting requests, lets current requests finish and then exits)

The mailing list might be a better forum for continued discussion.

Best,
David

from nameko.

Thank you so much for the reply.
I will use the mailing list from now on for any questions.
Cheers

from nameko.