Does anyone have implemented/Solution for Multi K8s cluster monitoring with Centralized Prometheus which is outside of the my K8S clusters.
Yes, have a look on federation chapter in the docs.
Or Thanos, Prometheus proxy or pushgateway
I am new to Prometheus.
Thank You for your reply do you have any reference doc or blog which has the detail.
Here is my requirement:
We have multiple Say (100 - 1000) K8s Cluster deployed with Kops with One Master and One Worker Node. The idea is to have a Centralized Monitoring (Prometheus) to monitor (100 - 1000) clusters.
Note: Each k8s cluster will be deployed in different VPC.
What i have implemented So for:
I could not find a better way to monitor multiple k8s Cluster deployed in different VPC’s hence we recommended to use the single VPC with different subnets.
I have my Proemtheus/Alertmanager/Grafana configured using docker compose and used the <kubernetes_sd_config> to scrape the metrics from multiple k8s Clusters.
The problem with this approach is we are not sure when a cluster gets created as the cluster creation by kops is automated when a new customer is onboard and getting the API Server Endpoint for each server is quite complicated when it comes to automation.
I am looking for the best approach in terms of monitoring multip (100-1000) that can be automated.
I think this project covers some of your bases, but i haven’t had a chance to play with it yet