Strange behavior in redis load balancing


(Dmitriy Shmatov) #1

I have current logging delivery flow:
Filebeat -> Redis -> Logstash -> Elasticsearch

Redis and Logstash are deployed in the same server. So I have array of (Redis + Logstash) servers and I whant to load balance my filebeat output requests to this servers array. Documentation says that I need to provide an array of IP's or hostnames and set loadbalance to true in redis config section on filebeat side. When I did it I got traffic jump on all my servers from 10 Mbit/s to 700 Mbit/s. But if I set loadbalance option to false (as Documentation says that server was chosen randomly from array) the traffic becomes normal to 10 Mbit/s. In both cases amout of logs were not changed. I have no clue what happens.

My filebeat configuration file:

filebeat.prospectors:
- input_type: log
  paths:
    - "/var/log/containers/*.log"
  document_type: container-logs
  symlinks: true
  tail_files: true
 
output.redis:
    hosts: ["10.0.0.1","10.0.0.2","10.0.0.3"]
    password: "some password"
    datatype: "list"
    key: "filebeat"
    db: 0
    loadbalance: true
    worker: 1

logging:
  level: error

(Steffen Siering) #2

how many logs do you have that will be shipped?

Have you tested max throughput for every single redis instance?

Have you had a look at filebeat log file?


(system) #3

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.