At first, I installed Thanos via the manifests directory of this repository (https://github.com/thanos-io/kube-thanos/tree/main/manifests) and everything worked fine. However, it seems to deploy not every component of Thanos. I cannot see the compactor and the query frontent for example.
Then I installed Thanos via the Helm Chart with the compactor and the query frontend as well, but the Helm Chart completely broke my kubernetes worker nodes because it ran into a OOM problem and extremely high load. In addition, the deployment looked a little bit different for the receiver for example. With kube-thanos the receiver is divided into in ingestor and a router and with the Helm Chart there is only one receiver component deployed.
I hope someone can help me, I would like to deploy all components including the compactor and query frontend without crashing my cluster.
kube-thanos is based on jsonnet and extends kube-prometheus. As part of kube-prometheus, the Prometheus Operator gets deployed, which is most likely the pieces you're missing.
Hello together,
I would like to know what is the difference between the Bitname Helm Chart of Thanos (https://artifacthub.io/packages/helm/bitnami/thanos) and Kube-Thanos?
At first, I installed Thanos via the manifests directory of this repository (https://github.com/thanos-io/kube-thanos/tree/main/manifests) and everything worked fine. However, it seems to deploy not every component of Thanos. I cannot see the compactor and the query frontent for example.
Then I installed Thanos via the Helm Chart with the compactor and the query frontend as well, but the Helm Chart completely broke my kubernetes worker nodes because it ran into a OOM problem and extremely high load. In addition, the deployment looked a little bit different for the receiver for example. With kube-thanos the receiver is divided into in ingestor and a router and with the Helm Chart there is only one receiver component deployed.
I hope someone can help me, I would like to deploy all components including the compactor and query frontend without crashing my cluster.