Release v1.12.0

Monitoring Core Module version 1.12.0

SIGHUP team maintains this module updated and tested. That is the main reason why we worked on this new release.
With the Kubernetes 1.21 release, it became the perfect time to start testing this module against this Kubernetes
release.

Continue reading the Changelog to discover them:

Changelog

• Update Grafana. From version 7.3.6 to 7.5.7.
• Update Prometheus Operator. From version 0.44.1 to 0.48.1.
• Update Prometheus. From version 2.22.2 to 2.27.1.
• Update thanos. From version 0.12.2 to 0.20.2.
• Update alertmanager. From version 0.21.0 to 0.22.2.
• Updte kube-state-metrics. From version 1.9.7 to 2.0.0.
  • Update all the Prometheus Rules and Dashboards.
• Update kube-proxy-metrics
• Update metrics-server. From version 0.4.1 to 0.5.0.
• Update node-exporter. From version 1.0.1 to 1.1.2.
• Update goldpinger. From version 3.0.0 to 3.2.0.
• Kubernetes support:
  • Deprecate Kubernetes 1.17 support.
  • Kubernetes 1.20 is considered stable.
  • Add tech-preview support to Kubernetes 1.21.
• Added x509-exporter package. Thanks to @nutellinoit
• Modify the alerts that track expiration of cluster certificates to fire within 30/7 days of expiration instead of 7/1 days. (kubeadm-k8s-rules, prometheus-k8s-rules)
• Add missing namespace field to configs/bases/default.
• Fixing thanos modules, missing namespace on components

Upgrade path

To upgrade this core module from v1.11.X to v1.12.0, first, delete old resources before applying the new
kustomize project, download this new version, then apply the kustomize project. No further action is required.

kustomize build katalog/prometheus-operator | kubectl apply -f -
kustomize build katalog/prometheus-operated | kubectl apply -f -
kustomize build katalog/grafana | kubectl apply -f -
kustomize build katalog/metrics-server | kubectl apply -f -
...

Release v1.11.1

Monitoring Core Module version v1.11.1

This patch contains only the change (and few examples) to move the container image from the reg.sighup.io registry
to registry.sighup.io.

Changelog

• All the container images comes from SIGHUP registry to avoid rate limits.

Upgrade path

To upgrade this core module from v1.11.0 to v1.11.1, you need to download this new version, then apply the
kustomize project. No further action is required.

kustomize build katalog/alertmanager-operated | kubectl apply -f -
kustomize build katalog/goldpinger | kubectl apply -f -
kustomize build katalog/grafana | kubectl apply -f -
kustomize build katalog/kube-proxy-metrics | kubectl apply -f -
kustomize build katalog/kube-state-metrics | kubectl apply -f -
kustomize build katalog/metrics-server | kubectl apply -f -
kustomize build katalog/node-exporter | kubectl apply -f -
kustomize build katalog/prometheus-operated | kubectl apply -f -
kustomize build katalog/prometheus-operator | kubectl apply -f -
kustomize build katalog/thanos | kubectl apply -f -

Release v1.11.0

Monitoring Core Module version 1.11.0

SIGHUP team maintains this module updated and tested. That is the main reason why we worked on this new release.
With the Kubernetes 1.20 release, it became the perfect time to start testing this module against this Kubernetes
release.

Continue reading the Changelog to discover them:

Changelog

• Update Prometheus Operator. From version 0.42.0 to 0.44.1.
• Update Prometheus. From version 2.21.0 to 2.22.2.
• Update Grafana. From version 7.1.5 to 7.3.6.
• Update metrics-server. From version 0.3.7 to 0.4.1.
• Kubernetes support:
  • Deprecate Kubernetes 1.16 support.
  • Kubernetes 1.19 is considered stable.
  • Add tech-preview support to Kubernetes 1.20.
• All the container images come from the SIGHUP registry to avoid rate limits.

Upgrade path

To upgrade this core module from v1.10.3 to v1.11.0, first, delete old resources before applying the new
kustomize project, download this new version, then apply the kustomize project. No further action is required.

kustomize build katalog/prometheus-operator | kubectl apply -f -
kustomize build katalog/prometheus-operated | kubectl apply -f -
kustomize build katalog/grafana | kubectl apply -f -
kustomize build katalog/metrics-server | kubectl apply -f -

Release v1.10.3

Monitoring Core Module version 1.10.3

kubeadm versions starting from 1.17 are not exposing kube-controller-manager nor kube-scheduler metrics
over the insecure port (respectively TCP/10252 and TCP/10251).
This makes the current ServiceMonitor configuration unable to properly scrape these targets.
Also, there were duplicated Prometheus Rules across packages: kubeadm-sm and prometheus-operated.

Thanks to @lnovara to spot and solve these issues.

Changelog

• FIX #51: kubeadm-sm package to proper configure serviceMonitors to solve kubeadm +1.17 versions.
• FIX: Remove duplicated prometheus rules:
  • KubeClientCertificateExpiration
  • KubeSchedulerDown
  • KubeControllerManagerDown

Upgrade path

To upgrade this core module from v1.10.2 to v1.10.3, you need to download this new version, then apply the
kustomize project. No further action is required.

kustomize build katalog/kubeadm-sm | kubectl apply -f -
kustomize build katalog/prometheus-operated | kubectl apply -f -

Release v1.10.2

Monitoring Core Module version 1.10.2

The previous 1.10.1 release fixes a couple of issues in the Fury alerting stack.
This release solves another problem, the NodeMachineIDCollision was deleted in 1.10.0.

Changelog

The previous 1.10.1 release includes a couple of issues in the Fury monitoring stack.
This release solves these problems to provide a smooth experience.

• coredns metrics service was in the default namespace instead of kube-system namespace. Moved
• In 1.10.0 monitoring module were updated but we forgot to update the dashboards. Dashboards updated.
• Add NodeMachineIDCollision as it was previously there.
• Add missing ovh-sm.

Upgrade path

To upgrade this core module from v1.10.1 to v1.10.2, you need to download this new version, then apply the
kustomize project. No further action is required.

kustomize build katalog/prometheus-operator | kubectl apply -f -
kustomize build katalog/prometheus-operated | kubectl apply -f -
kustomize build katalog/node-exporter | kubectl apply -f -
kustomize build katalog/metrics-server | kubectl apply -f -
kustomize build katalog/kubeadm-sm | kubectl apply -f -
kustomize build katalog/kube-state-metrics | kubectl apply -f -
kustomize build katalog/kube-proxy-metrics | kubectl apply -f -
kustomize build katalog/grafana | kubectl apply -f -
kustomize build katalog/goldpinger | kubectl apply -f -
kustomize build katalog/gke-sm | kubectl apply -f -
kustomize build katalog/eks-sm | kubectl apply -f -
kustomize build katalog/alertmanager-operated | kubectl apply -f -
kustomize build katalog/aks-sm | kubectl apply -f -
kustomize build katalog/ovh-sm | kubectl apply -f -

Then, remove the buggy coredns service and its associated servicemonitor:

kubectl delete service -n default coredns
kubectl delete servicemonitor -n default coredns
kubectl delete configmap -n default sighup-coredns-dashboard-definition

Release v1.10.1

Monitoring Core Module version 1.10.1

The previous 1.10.0 release includes a couple of issues in the Fury alerting stack.
This release solves these problems to maintain retro compatibility with previous alerting stack versions.

Special thanks to @lnovara, @nutellinoit, and @lzecca78 to spot these problems!

Changelog

• Remove CPUThrottlingHigh as it was previously removed in #43.
• Rename Watchdog to DeadMansSwitch to maintain backward compatibility.

Upgrade path

To upgrade this core module from v1.10.0 to v1.10.1, you need to download this new version, then apply the
kustomize project. No further action is required.

kustomize build katalog/prometheus-operated | kubectl apply -f -

Release v1.10.0

Monitoring Core Module version 1.10.0

SIGHUP team maintains this module updated and tested. That is the main reason why we worked on this new release.
With the Kubernetes 1.19 release, it became the perfect time to start testing this module against this Kubernetes
release. Grafana Labs released a major release of its flagship product, Grafana 7. SIGHUP includes the latest Grafana
release in this Fury distribution core module. Alongside this major change, an extra interesting job has been developed.
Continue reading the Changelog to discover them:

Changelog

• Update Prometheus Operator. From version 0.37.0 to 0.42.0.
• Update Prometheus. From version 2.16.0 to 2.21.0.
• Refactor infrastructure dependant service monitors. Thanks to @lnovara!
  • Follow the upgrade path section to understand how to deploy these changes.
• Update Alertmanager. From version 0.20.0 to 0.21.0.
• Update Grafana. From version 6.6.2 to 7.1.5.
  • Decoupled pre-defined dashboards.
  • Add the internal Grafana metrics dashboard.
• Update node exporter. From version 0.18.1 to 1.0.1.
• Update kube-state-metrics. From version 1.9.5 to 1.9.7.
• Update metrics-server. From version 0.3.6 to 0.3.7.
• Update Goldpinger. From version 2.0.0 to 3.0.0.
• Add kube-proxy-metrics package.
• Add eks-sm package.
• Remove kong and glusterfs. If you need to recover them, take a look at the
  1.9.0 version of this module.

Upgrade path

This release includes a deep refactoring of the Grafana dashboards and Prometheus Operator Service Monitors objects.
Moved from all in one deployment (all dashboards were shipped by default) to just deploy those dashboards having
packages associated.
As an example, if you won't deploy goldpinger
you will not have its Grafana dashboard nor Prometheus Operator Service Monitor deployed.

To upgrade this core module from v1.9.0 to v1.10.0, first, delete old resources before applying the new
kustomize project.

$ kubectl delete cm -A -l grafana-sighup-dashboard=default
configmap "grafana-dashboard-apiserver-h6d8ghht52" deleted
configmap "grafana-dashboard-cluster-total-tgghggcfc5" deleted
configmap "grafana-dashboard-controller-manager-bf5555kfb7" deleted
configmap "grafana-dashboard-definitions-gatekeeper-ttb7264bk7" deleted
configmap "grafana-dashboard-definitions-gluster-tdh82ht56d" deleted
configmap "grafana-dashboard-definitions-goldpinger-6ffd8hg8mm" deleted
configmap "grafana-dashboard-definitions-k8s-4bdf22m55f" deleted
configmap "grafana-dashboard-definitions-kong-tf4m544522" deleted
configmap "grafana-dashboard-definitions-nginx-657dtk9948" deleted
configmap "grafana-dashboard-k8s-resources-cluster-fbfdmt68gg" deleted
configmap "grafana-dashboard-k8s-resources-namespace-46654778gf" deleted
configmap "grafana-dashboard-k8s-resources-node-bg8fddhkb2" deleted
configmap "grafana-dashboard-k8s-resources-pod-hg47c4fmf6" deleted
configmap "grafana-dashboard-k8s-resources-workload-d7t5t5tt45" deleted
configmap "grafana-dashboard-k8s-resources-workloads-namespace-865f6f8k97" deleted
configmap "grafana-dashboard-kubelet-h5b284cf8c" deleted
configmap "grafana-dashboard-namespace-by-pod-7hckk9mmch" deleted
configmap "grafana-dashboard-namespace-by-workload-mc98g9c5ft" deleted
configmap "grafana-dashboard-node-cluster-rsrc-use-8hkc5dk64m" deleted
configmap "grafana-dashboard-node-rsrc-use-h46hdmfd62" deleted
configmap "grafana-dashboard-nodes-88kc4mmfh5" deleted
configmap "grafana-dashboard-persistentvolumesusage-hh2h2b9k5f" deleted
configmap "grafana-dashboard-pod-total-49529c5cff" deleted
configmap "grafana-dashboard-pods-dbk2h5d9c6" deleted
configmap "grafana-dashboard-prometheus-gffb4c9m9m" deleted
configmap "grafana-dashboard-prometheus-remote-write-8bmfkf2d65" deleted
configmap "grafana-dashboard-proxy-2mmgm84b5c" deleted
configmap "grafana-dashboard-scheduler-7b2dkmc28d" deleted
configmap "grafana-dashboard-statefulset-8ghkb269k6" deleted
configmap "grafana-dashboard-workload-total-44d5hd6g22" deleted
$ kubectl delete servicemonitor -n monitoring alertmanager goldpinger grafana kube-state-metrics node-exporter prometheus prometheus-operator
servicemonitor.monitoring.coreos.com "alertmanager" deleted
servicemonitor.monitoring.coreos.com "goldpinger" deleted
servicemonitor.monitoring.coreos.com "grafana" deleted
servicemonitor.monitoring.coreos.com "kube-state-metrics" deleted
servicemonitor.monitoring.coreos.com "node-exporter" deleted
servicemonitor.monitoring.coreos.com "prometheus" deleted
servicemonitor.monitoring.coreos.com "prometheus-operator" deleted
$ kubectl delete servicemonitor -n kube-system coredns etcd-metrics kube-apiserver kube-controller-manager kube-scheduler kubelet
servicemonitor.monitoring.coreos.com "coredns" deleted
servicemonitor.monitoring.coreos.com "etcd-metrics" deleted
servicemonitor.monitoring.coreos.com "kube-apiserver" deleted
servicemonitor.monitoring.coreos.com "kube-controller-manager" deleted
servicemonitor.monitoring.coreos.com "kube-scheduler" deleted
servicemonitor.monitoring.coreos.com "kubelet" deleted

Finally, download this new version, then apply the kustomize project. No further action is required.

kustomize build katalog/calico | kubectl apply -f -

Important note to Furyctl users

If you are using furyctl and Furyfiles.yml
you have to add a new entry in your Furyfile.yml file while deploying
Kubernetes platform servicemonitors like kubeadm-sm, gke-sm...:

Before:

bases:
  - name: monitoring/kubeadm-sm # Same with gke-sm, aks-sm, kubeadm-sm
    version: v1.9.0

After:

bases:
  - name: monitoring/kubeadm-sm # Same with gke-sm, aks-sm, kubeadm-sm
    version: v1.10.0
  - name: monitoring/configs # Same with gke-sm, aks-sm, kubeadm-sm
    version: v1.10.0