[Request] environment variable KUROKO2_WORKER=$worker_id for jobs

[FUJI Goro]

Sometimes I'd like to know the process is running on workers or not.

For example, when I configure arproxy to report slow queries to Sentry, the threshold in workers (i.e. OLAP) differs from that in application servers (i.e. OLTP).

Of course I can define that environment variables in each job, but if this idea is worth considering, I'll make a pull request for it.

What do you think of it?

[Kohei Suzuki]

Let me confirm your situation.

• You have 2 (or more) kind of servers (e.g., worker-medium, worker-powerful)
• You have a job X which uses SOME_THRESHOLD environment variable
• SOME_THRESHOLD depends on the worker kind
  • SOME_THRESHOLD should be 5 in worker-medium
  • SOME_THRESHOLD should be 1 in worker-powerful
• You're thinking of adding KUROKO2_WORKER variable and deducing worker kind from it

Is it right?

[FUJI Goro]

Ah, in fact, currently I have just single worker; I'd like to distinguish workers from application servers. That is, I won't use $worker_id for now.

[Sorah Fukumori]

Why you can't configure a environment variable for kuroko2 process in your worker servers?

[FUJI Goro]

Of course I can configure so, but I think it might be a general idea that is worth considering.

Feel free to close this issue if you think it's not so useful 😄

[Kohei Suzuki]

Please re-open when you need it. I still don't understand the use case.