NETOBSERV-547 Tweak loki config

[jpinsonneau]

Here is a non exhaustive list of errors with related tweaks you may apply.

We can try to better define some default values according to cluster sizes or flow processed per second in a table format in followup.

[memodi]

/cc @nathan-weinberg

LGTM

[nathan-weinberg]

Thanks for this @jpinsonneau - please add details as to how/where to change these configs (i.e. the kubectl or oc commands) as well

[eranra]

@nathan-weinberg this answers a lot of the questions around (run without limits)

@jpinsonneau this is great info

[eranra]

@jpinsonneau what about automation of all of those ... can we somehow "do" that --- maybe create a script to run over the operator that we add in the instructions for customers?

[jpinsonneau]

@jpinsonneau what about automation of all of those ... can we somehow "do" that --- maybe create a script to run over the operator that we add in the instructions for customers?

For loki simple / distributed it's easy to integrate these in the provided config. I'll double check these but we are in a good shape for now.
We decided to take more time for loki-operator as it needs parameters defined in their CRD. I wonder if we can do something more flexible to allow any config without having a huge CRD on their side.

[jpinsonneau]

Thanks for this @jpinsonneau - please add details as to how/where to change these configs (i.e. the kubectl or oc commands) as well

Sure I can document this for each situations: simple / distributed / loki-operator

[jpinsonneau]

Thanks for this @jpinsonneau - please add details as to how/where to change these configs (i.e. the kubectl or oc commands) as well

Here we go dee092b @nathan-weinberg

[nathan-weinberg]

/lgtm