Customized Logs & Optimize Infrastructure

Hi, i deployed a ManageIQ Infrastructure with two appliances by zone in two different zones and one ManageIQ DB with a replication.

So now i wonder how dispatch servers roles, workers and how customized logs settings on my appliances (4 CPU & 12GB RAM) to balance memory usage on them.
So if someone can teach me goods practices, i’m grateful.

I’m very interested by send my logs in a solution like ELK with a JSON format

Hi Gregoire

Did you manage to have a look at the “Deploying CloudForms at Scale” reference architecture: Deploying CloudForms at Scale Reference Architectures 2017 | Red Hat Customer Portal ?

If you enable workers on multiple appliances in a zone, the workers should auto-balance. Some worker types can run multiple instances in a zone, whereas other worker types can only have a single primary copy active, with others in secondary or standby mode.

You may need more than 12GB memory in your appliances, I’d normally start with 16GB as a minimum and work upwards from there, depending on what your MIQ infrastructure is doing.

There’s an interesting discussion of logging in this thread:

Hope this helps,
pemcg

Hi @pemcg

I’m gonna go through them thank you !

About customized log i want to be able to change the directory and turn them into JSON file.

@Fryguy didn’t you implement something like this for common logging?

1 Like

I did but the output is only on the podified architecture. There might be a way to enable JSON logging as an option for appliances. That being said, on appliances we log to journald, so if the logging stack can read journald logs then you should be able to ingest them.

Hi @Fryguy, thanks for your answer, i will go look into it.

Can you tell me if i can install a filebeat client on my appliance without impact the appliance ?

I’m not sure about filebeat specifically, as I’ve not heard of others who’ve used it, but in my experience adding a monitoring agent in general doesn’t usually impact the appliance, so I say go for it.

1 Like