.ops: add ecamp3-logging
Finally i have a branch again which works. Now with a lot less yaml.
This adds the EFFK (Elasticsearch, fluentbit, fluentd, Kibana) stack to the ops folder. Kibana and Elasticsearch taken from https://www.digitalocean.com/community/tutorials/how-to-set-up-an-elasticsearch-fluentd-and-kibana-efk-logging-stack-on-kubernetes
For the full story, see: https://github.com/ecamp/ecamp3/issues/4775
Nice, thank you @BacLuc
I was not that deeply involved in the discussion in #4775. So this is now a mix between option 1 and 2? Do you have an estimation, how much additional nodes we'd need to run it? Is this deployed to dev already to test out?
I was not that deeply involved in the discussion in #4775. So this is now a mix between option 1 and 2?
Yes, but with the fluent operator and fluentbit and fluentd instead of filebeat. You see the list of the tried options here: https://github.com/ecamp/ecamp3/issues/4775#issuecomment-2080457061 This is the last one
Do you have an estimation, how much additional nodes we'd need to run it?
Currently on dev it looks like this:
| NAME | CPU(cores) | MEMORY(bytes) |
|---|---|---|
| elasticsearch-cluster-0 | 18m | 977Mi |
| fluent-bit-4mh29 | 4m | 56Mi |
| fluent-bit-b92zd | 3m | 52Mi |
| fluent-bit-fdmzw | 3m | 47Mi |
| fluent-bit-s5rgz | 4m | 48Mi |
| fluent-operator-7c5d78987c-tbx6l | 1m | 49Mi |
| fluentd-0 | 6m | 155Mi |
| kibana-7cc44b6d66-jllsq | 86m | 523Mi |
| Total | 1907Mi |
So we would need another node. Maybe 2 if elasticsearch contains a lot of logs, we will see.
Is this deployed to dev already to test out?
Yes, you can do a port forward with
k port-forward services/kibana 5601:5601