Explanation why a seperate run task is necessary and usage of sidecar

[h0jeZvgoxFepBQ2C]

Hi,

since no "Discussion" tab is available, i allowed myself to open an issue, since this could maybe be added to the Readme:

I have 2 questions:

1) why is a seperate process necessary - and not just an engine which is mounted under a specific route (like f.e. the sidekiq web interface). I think the stats could be delivered by the rails process itself, or?

2) is it recommended to use the task as a sidecar with the rails process pods? We want to save pods amounts (due to f.e. hard limit of 110 pods per node), and maybe a sidecar would help here?

Thank you!

[SamSaffron]

separate process is not necessary, but having 10 metric endpoints for 10 unicorn workers would be a nightmare to manage and distributed storage is enormous amounts of pointless work.

pods work / extra process works / if you just have a single pid being monitored in the container then integrated can work.