`max_workers` when starting multiple services

googlegroup
1

Hi!

Still loving Nameko - So many problems, solved so quickly!

One question about the `max_workers` config.

If I run multiple instances on the Nameko runner, for example:

nameko run --config=config.yaml service:ServiceA
nameko run --config=config.yaml service:ServiceB
nameko run --config=config.yaml service:ServiceC

I expect that they will each get thier own pool of workers up to
`max_workers`.

But what happens if I run a single instance, and let Nameko register the
individual services? For example:

nameko run --config=config.yaml service

My guess is that the `max_workers` pool size is shared between all the
services, but I thought I'd just ask the question :slight_smile:

Many thanks,

Geoff

2

Hi Geoff,

Each service runs in its own independent container which is where limits
like max_workers are set, so each service would still have its own worker
pool

Best,
David

···

On Friday, 27 October 2017 18:10:19 UTC+1, juko...@gmail.com wrote:

Hi!

Still loving Nameko - So many problems, solved so quickly!

One question about the `max_workers` config.

If I run multiple instances on the Nameko runner, for example:

nameko run --config=config.yaml service:ServiceA
nameko run --config=config.yaml service:ServiceB
nameko run --config=config.yaml service:ServiceC

I expect that they will each get thier own pool of workers up to
`max_workers`.

But what happens if I run a single instance, and let Nameko register the
individual services? For example:

nameko run --config=config.yaml service

My guess is that the `max_workers` pool size is shared between all the
services, but I thought I'd just ask the question :slight_smile:

Many thanks,

Geoff

3

That is awesome news. I have been running services in separate Supervisord
processed :slight_smile: But now I'm running more and more microservices, the
supervisord lists are getting out of hand :slight_smile:

Thanks for the confirmation David.

···

On Friday, October 27, 2017 at 11:33:22 AM UTC-7, David Szotten wrote:

Hi Geoff,

Each service runs in its own independent container which is where limits
like max_workers are set, so each service would still have its own worker
pool

Best,
David

On Friday, 27 October 2017 18:10:19 UTC+1, juko...@gmail.com wrote:

Hi!

Still loving Nameko - So many problems, solved so quickly!

One question about the `max_workers` config.

If I run multiple instances on the Nameko runner, for example:

nameko run --config=config.yaml service:ServiceA
nameko run --config=config.yaml service:ServiceB
nameko run --config=config.yaml service:ServiceC

I expect that they will each get thier own pool of workers up to
`max_workers`.

But what happens if I run a single instance, and let Nameko register the
individual services? For example:

nameko run --config=config.yaml service

My guess is that the `max_workers` pool size is shared between all the
services, but I thought I'd just ask the question :slight_smile:

Many thanks,

Geoff

4

np

note that there are other reasons you might want to run separate processes,
e.g. limiting the impact of something accidentally hogging the cpu and
blocking the event loop

best,
d

···

On Friday, 27 October 2017 19:44:33 UTC+1, juko...@gmail.com wrote:

That is awesome news. I have been running services in separate Supervisord
processed :slight_smile: But now I'm running more and more microservices, the
supervisord lists are getting out of hand :slight_smile:

Thanks for the confirmation David.

On Friday, October 27, 2017 at 11:33:22 AM UTC-7, David Szotten wrote:

Hi Geoff,

Each service runs in its own independent container which is where limits
like max_workers are set, so each service would still have its own worker
pool

Best,
David

On Friday, 27 October 2017 18:10:19 UTC+1, juko...@gmail.com wrote:

Hi!

Still loving Nameko - So many problems, solved so quickly!

One question about the `max_workers` config.

If I run multiple instances on the Nameko runner, for example:

nameko run --config=config.yaml service:ServiceA
nameko run --config=config.yaml service:ServiceB
nameko run --config=config.yaml service:ServiceC

I expect that they will each get thier own pool of workers up to
`max_workers`.

But what happens if I run a single instance, and let Nameko register the
individual services? For example:

nameko run --config=config.yaml service

My guess is that the `max_workers` pool size is shared between all the
services, but I thought I'd just ask the question :slight_smile:

Many thanks,

Geoff

5

Nearly 3 years later and this reply helped me solve a performance issue I was having. Thanks again!