feat: adding casspoke

.gitignore

@@ -1,2 +1,3 @@
 .idea
 tests/.vagrant
+dind-cluster-*

.travis.yml

@@ -3,7 +3,6 @@ language: go
 sudo: required
 
 env:
-  - K8S_VERSION=1.9 HELM_VERSION=2.9.1 NUM_NODES=3 SKIP_SNAPSHOT=y PATH="$HOME/.kubeadm-dind-cluster:/tmp/linux-amd64:$PATH"
   - K8S_VERSION=1.8 HELM_VERSION=2.9.1 NUM_NODES=3 SKIP_SNAPSHOT=y PATH="$HOME/.kubeadm-dind-cluster:/tmp/linux-amd64:$PATH"
 
 install:
@@ -17,4 +16,4 @@ script:
   - bats -t tests/states/delete.bats
 
 after_script:
-  - tests/k8s-euft/clean_k8s.sh
+  - tests/k8s-euft/clean_k8s.sh

README.md

@@ -2,7 +2,7 @@
 
 You can find here a helm chart we're using at [MySocialApp](https://mysocialapp.io)
 
-This is a Kubernetes Helm Chart for Cassandra.
+This is a Kubernetes Helm Chart for Cassandra with several useful monitoring and management tools.
 
 ## Backups
 
@@ -94,6 +94,18 @@ cassandraAlertmanager:
 
 And adapt alert labels to you configuration.
 
+## Casspoke
+
+[Casspoke](https://github.com/criteo/casspoke) is a latency probe checker for Cassandra. To enable it:
+
+```yaml
+# Casspoke
+casspoke:
+  enablePoke: true
+```
+
+Update as well other settings if you need.
+
 # FAQ
 ## How to replace a Node?
 

kubernetes/templates/cassPokeConfigmap.yaml

@@ -0,0 +1,27 @@
+{{- if .Values.casspoke.enablePoke }}
+kind: ConfigMap
+apiVersion: v1
+metadata:
+  name: casspoke
+  namespace: {{ .Release.Namespace }}
+  labels:
+    app: casspoke
+    chart: {{ .Chart.Name }}-{{ .Chart.Version | replace "+" "_" }}
+    release: {{ .Release.Name }}
+    heritage: {{ .Release.Service }}
+data:
+  config.yml: |-
+    app:
+      measurementPeriodInSec: 10
+      refreshDiscoveryPeriodInSec: 60
+      httpServerPort: {{ .Values.casspoke.config.httpServerPort }}
+
+    discovery:
+      staticDns:
+        clustername: {{ .Values.casspoke.config.clustername }}
+        host: {{ .Values.casspoke.config.host }}
+
+    service:
+      type: CassandraRunner
+      timeoutInSec: 60
+{{- end }}

kubernetes/templates/casspokeDeploy.yaml

@@ -0,0 +1,51 @@
+{{- if .Values.casspoke.enablePoke }}
+apiVersion: extensions/v1beta1
+kind: Deployment
+metadata:
+  name: casspoke
+  namespace: {{ .Release.Namespace }}
+  labels:
+    app: casspoke
+    chart: {{ .Chart.Name }}-{{ .Chart.Version | replace "+" "_" }}
+    release: {{ .Release.Name }}
+    heritage: {{ .Release.Service }}
+spec:
+  replicas: {{ .Values.casspoke.replicaCount }}
+  strategy:
+    type: RollingUpdate
+    rollingUpdate:
+      maxSurge: 1
+  template:
+    metadata:
+      labels:
+        app: casspoke
+    spec:
+      nodeSelector:
+        {{- range $key, $val := .Values.casspoke.nodeSelector }}
+        {{ $key }}: {{ $val | quote }}
+        {{- end}}
+      containers:
+      - name: casspoke
+        image: "quay.io/deimosfr/casspoke:{{ .Values.casspoke.imageVersion }}"
+        env:
+        - name: JVM_OPTS
+          value: {{ .Values.casspoke.envVariables.JVM_OPTS }}
+        ports:
+        - name: metrics
+          containerPort: {{ .Values.casspoke.config.httpServerPort }}
+          protocol: TCP
+        livenessProbe:
+          tcpSocket:
+            port: {{ .Values.casspoke.config.httpServerPort }}
+          initialDelaySeconds: 60
+          periodSeconds: 10
+        resources:
+{{ toYaml .Values.casspoke.resources | indent 10 }}
+        volumeMounts:
+          - name: casspoke
+            mountPath: /etc/casspoke
+      volumes:
+        - name: casspoke
+          configMap:
+            name: casspoke
+{{- end }}

kubernetes/templates/casspokeService.yaml

@@ -0,0 +1,21 @@
+{{- if .Values.casspoke.enablePoke }}
+apiVersion: v1
+kind: Service
+metadata:
+  name: casspoke
+  namespace: {{ .Release.Namespace }}
+  labels:
+    app: casspoke
+    chart: {{ .Chart.Name }}-{{ .Chart.Version | replace "+" "_" }}
+    release: {{ .Release.Name }}
+    heritage: {{ .Release.Service }}
+spec:
+  clusterIP: None
+  ports:
+    - name: metrics
+      port: 80
+      targetPort: {{ .Values.casspoke.config.httpServerPort }}
+      protocol: TCP
+  selector:
+    app: casspoke
+{{- end }}

kubernetes/templates/casspokeServicemonitor.yaml

@@ -0,0 +1,21 @@
+{{- if .Values.casspoke.enablePoke }}
+apiVersion: monitoring.coreos.com/v1
+kind: ServiceMonitor
+metadata:
+  name: casspoke-{{ .Release.Namespace }}
+  namespace: observability
+  labels:
+    app: casspoke
+    prometheus: observability
+spec:
+  selector:
+    matchLabels:
+      app: casspoke
+  namespaceSelector:
+    matchNames:
+      - {{ .Release.Namespace }}
+  endpoints:
+  - port: metrics
+    interval: 60s
+    scrapeTimeout: 59s
+{{- end }}

kubernetes/values.yaml

@@ -66,6 +66,27 @@ cassandraExporter:
     listenPort: 8080
     jvmOpts: ""
 
+# Casspoke
+casspoke:
+  enablePoke: false
+  imageVersion: 1.2
+  replicaCount: 3
+  nodeSelector:
+    node-role.kubernetes.io/node: "true"
+  config:
+    httpServerPort: 8080
+    clustername: cassandra
+    host: cassandra-0.cassandra:9042,cassandra-1.cassandra:9042,cassandra-2.cassandra:9042
+  envVariables:
+    JVM_OPTS: "-XX:+UnlockExperimentalVMOptions -XX:MaxRAMFraction=2 -XX:+AlwaysPreTouch"
+  resources:
+    limits:
+      cpu: 300m
+      memory: 500Mi
+    requests:
+      cpu: 300m
+      memory: 500Mi
+
 # Prometheus scraping
 cassandraPrometheusScrap:
   enableScrap: false

tests/states/deploy.bats

@@ -35,3 +35,7 @@ load common
 @test "Check all Cassandra nodes are Up and Normal" {
     num_nodes_eq_nodetool_status_un
 }
+
+@test "Upgrade test" {
+    helm upgrade kubernetes -n cassandra
+}