K
Q

Kubernetes logging architecture - NFS Persistent volume?

5/12/2019

In kubernetes cluster for microservice application to diagnose the issue we need logs.

  1. Is is good idea to use NFS persistent volume for all microservice logs?
  2. If yes, Is it possible to apply log rotation policy on NFS persistent volume based on size or days?
  3. If we use ELK stack with filebeat it will need more resources and learning for customer to get the required log.

What will be best approach i.e NFS or ELK stack or mixed?

-- Mr.Pramod Anarase
kubernetes
logging

Similar Questions

how to write daily backup job for elasticsearch in production environment, we are using docker image and kubernetes for containerized platform
How to configure docker CLI to access docker server from minikube on Git-Bash?
Installation of Istio on GKE / Google Cloud
Istio - Connect to an external ip
How to add a default value to a kubernetes helm template
How to create critical section in nodejs multi-instance application in Kubernetes?
Can somebody explain whay I have to use external (MetalLB, HAProxy etc) Load Balancer with Bare-metal kubernetes cluster?
Is displaying/printing a "Hello World" message in Java/Python same as in Kubernetes?
Exporting spark executor jmx metrics for multiple executors running in same machine
Enable autostart of kubernetes service
Message "error: error parsing mongodb-deployment.yaml: error converting YAML to JSON: yaml: line 27: did not find expected key"
Argo Rollout Dependency Ordering
Spark driver doesn't crash on exception
How I get the number of (currently) established TCP connections in prometheus (kubernetes monitoring)

1 Answer

5/13/2019
  1. NFS is ok as long as it is able to offer required performance.
  2. You should apply lifecycle policy at Elasticsearch indices level. Modern Kibana has a nice interface for creation of lifecycle policies and overall monitoring of ES.
  3. Never worked with Filebeat. We use EFK stack - Elasticsearch, Fluentd and Kibana. It works pretty well and is installed only using Helm Charts.
-- Vasily Angapov
Source: StackOverflow