Implementing DaemonSet for Telegraf

fmdlc · Oct 23, 2020 · de2d87e · de2d87e
1 parent cc7c92e
commit de2d87e
Show file tree

Hide file tree

Showing 6 changed files with 237 additions and 134 deletions.
diff --git a/.pre-commit-config.yaml b/.pre-commit-config.yaml
@@ -5,6 +5,7 @@ repos:
       - id: check-symlinks
       - id: check-yaml
         exclude: kubernetes/yaml/network-dashboard.yaml|kubernetes/ISP-Checker-deploy.yaml
+        args: ["--allow-multiple-documents"]
       - id: detect-aws-credentials
       - id: detect-private-key
       - id: double-quote-string-fixer

diff --git a/kubernetes/ISP-Checker-deploy.yaml b/kubernetes/ISP-Checker-deploy.yaml
@@ -33,6 +33,48 @@ spec:
       storage: 1Gi
 ---
 apiVersion: v1
+kind: Secret
+metadata:
+  namespace: monitoring
+  name: isp-checker-secrets-rw
+type: Opaque
+stringData:
+  INFLUXDB_DB: telegraf
+  INFLUXDB_ADMIN_USER: admin
+  INFLUXDB_ADMIN_PASSWORD: VerySecurePassword
+
+---
+apiVersion: v1
+kind: Secret
+metadata:
+  namespace: monitoring
+  name: isp-checker-secrets-ro
+type: Opaque
+stringData:
+  INFLUXDB_DB: telegraf
+  INFLUXDB_READ_USER: grafana
+  INFLUXDB_READ_PASSWORD: VerySecurePassword
+
+---
+apiVersion: v1
+kind: Secret
+metadata:
+  namespace: monitoring
+  name: grafana-datasource
+stringData:
+  datasource.yaml: |-
+    datasources:
+      - name: InfluxDB
+        type: influxdb
+        access: proxy
+        orgId: 1
+        url: http://influxdb-svc.monitoring.svc.cluster.local:8086/
+        user: grafana
+        database: telegraf
+        isDefault: true
+        editable: true
+---
+apiVersion: v1
 kind: ConfigMap
 metadata:
   namespace: monitoring
@@ -74,17 +116,18 @@ data:
   telegraf.conf: |+
     [global_tags]
     [agent]
-      interval = "10s"
+      interval = "60s"
       round_interval = true
       metric_batch_size = 1000
       metric_buffer_limit = 10000
       collection_jitter = "0s"
       flush_interval = "10s"
       flush_jitter = "0s"
       precision = ""
-      hostname = ""
+      hostname = "$HOSTNAME"
       omit_hostname = false
       debug = false
+      quiet = true
 
     [[outputs.influxdb]]
       urls = ["http://influxdb-svc.monitoring.svc.cluster.local:8086/"]
@@ -95,30 +138,38 @@ data:
       user_agent = "telegraf"
       skip_database_creation = false
 
+    [[inputs.net]]
+    [[inputs.netstat]]
+    [[inputs.diskio]]
+    [[inputs.linux_sysctl_fs]]
+    [[inputs.mem]]
+    [[inputs.processes]]
+    [[inputs.swap]]
+    [[inputs.system]]
+
     [[inputs.cpu]]
       percpu = true
       totalcpu = true
       collect_cpu_time = false
       report_active = false
 
     [[inputs.disk]]
-      ignore_fs = ["tmpfs", "devtmpfs", "devfs", "iso9660", "overlay", "aufs", "squashfs"]
-
-    [[inputs.diskio]]
-
-    [[inputs.kernel]]
-
-    [[inputs.mem]]
-
-    [[inputs.processes]]
-
-    [[inputs.swap]]
-
-    [[inputs.system]]
+      ignore_fs = ["tmpfs",
+        "devtmpfs",
+        "devfs",
+        "iso9660",
+        "overlay",
+        "aufs",
+        "squashfs"
+      ]
 
     [[inputs.dns_query]]
        servers = ["4.2.2.1", "8.8.8.8"]
-       domains = ["www.google.com", "www.twitter.com", "www.amazon.com", "www.wikipedia.org"]
+       domains = ["www.google.com",
+         "www.twitter.com",
+         "www.amazon.com",
+         "www.wikipedia.org"
+       ]
        record_type = "A"
        port = 53
        timeout = 3
@@ -137,12 +188,6 @@ data:
     [[inputs.internal]]
       collect_memstats = true
 
-    [[inputs.linux_sysctl_fs]]
-
-    [[inputs.net]]
-
-    [[inputs.netstat]]
-
     [[inputs.ping]]
       urls = [
           "google.com",
@@ -174,51 +219,22 @@ data:
       timeout = "3m"
       interval = "15m"
       data_format = "csv"
-      csv_column_names = ['Server ID','Sponsor','Server Name','Timestamp','Distance','Ping','Download','Upload','Share','IP Address']
+      csv_column_names = ['Server ID',
+        'Sponsor',
+        'Server Name',
+        'Timestamp',
+        'Distance',
+        'Ping',
+        'Download',
+        'Upload',
+        'Share',
+        'IP Address'
+      ]
       csv_timestamp_column = "Timestamp"
       csv_timestamp_format = "2006-01-02T15:04:05Z07:00"
----
-apiVersion: v1
-kind: Secret
-metadata:
-  namespace: monitoring
-  name: isp-checker-secrets-rw
-type: Opaque
-stringData:
-  INFLUXDB_DB: telegraf
-  INFLUXDB_ADMIN_USER: admin
-  INFLUXDB_ADMIN_PASSWORD: VerySecurePassword
-
----
-apiVersion: v1
-kind: Secret
-metadata:
-  namespace: monitoring
-  name: isp-checker-secrets-ro
-type: Opaque
-stringData:
-  INFLUXDB_DB: telegraf
-  INFLUXDB_READ_USER: grafana
-  INFLUXDB_READ_PASSWORD: VerySecurePassword
 
----
-apiVersion: v1
-kind: Secret
-metadata:
-  namespace: monitoring
-  name: grafana-datasource
-stringData:
-  datasource.yaml: |-
-    datasources:
-      - name: InfluxDB
-        type: influxdb
-        access: proxy
-        orgId: 1
-        url: http://influxdb-svc.monitoring.svc.cluster.local:8086/
-        user: grafana
-        database: telegraf
-        isDefault: true
-        editable: true
+    [[inputs.docker]]
+      endpoint = "unix:///var/run/docker.sock"
 ---
 apiVersion: v1
 kind: ConfigMap
@@ -3983,33 +3999,55 @@ spec:
             secretName: grafana-datasource
 ---
 apiVersion: apps/v1
-kind: Deployment
+kind: DaemonSet
 metadata:
   namespace: monitoring
   name: telegraf
   labels:
     app: telegraf
 spec:
-  replicas: 1
   selector:
     matchLabels:
       app: telegraf
   template:
     metadata:
       labels:
+        name: telegraf
         app: telegraf
     spec:
       containers:
         - name: telegraf
           image: tty0/isp_telegraf:1.15.3
+          imagePullPolicy: IfNotPresent
           envFrom:
             - secretRef:
                 name: isp-checker-secrets-rw
-          imagePullPolicy: IfNotPresent
+          env:
+            - name: "KUBE_API"
+              value: "https://node2:6443/"
+            - name: HOSTNAME
+              valueFrom:
+                fieldRef:
+                  fieldPath: spec.nodeName
+            - name: "HOST_PROC"
+              value: "/rootfs/proc"
+            - name: "HOST_SYS"
+              value: "/rootfs/sys"
           volumeMounts:
             - name: telegraf-config-volume
               mountPath: /etc/telegraf/
               readOnly: true
+            - name: sys
+              mountPath: /sys/
+              readOnly: true
+            - name: proc
+              mountPath: /proc/
+              readOnly: true
+            - name: docker-socket
+              mountPath: /var/run/docker.sock
+            - name: utmp
+              mountPath: /var/run/utmp
+              readOnly: true
           resources:
             requests:
               memory: "100Mi"
@@ -4021,7 +4059,20 @@ spec:
         - name: wait-for-influxdb
           image: curlimages/curl:7.72.0
           command: ['sh', '-c', "until curl -XGET http://influxdb-svc.$(cat /var/run/secrets/kubernetes.io/serviceaccount/namespace).svc.cluster.local:8086/; do echo Waiting\ for\ InfluxDB; sleep 2; done"]
+      terminationGracePeriodSeconds: 30
       volumes:
         - name: telegraf-config-volume
           configMap:
             name: telegraf-config
+        - name: sys
+          hostPath:
+            path: /sys
+        - name: docker-socket
+          hostPath:
+            path: /var/run/docker.sock
+        - name: proc
+          hostPath:
+            path: /proc
+        - name: utmp
+          hostPath:
+            path: /var/run/utmp
diff --git a/kubernetes/yaml/configmap.yaml b/kubernetes/yaml/configmap.yaml
@@ -41,17 +41,18 @@ data:
   telegraf.conf: |+
     [global_tags]
     [agent]
-      interval = "10s"
+      interval = "60s"
       round_interval = true
       metric_batch_size = 1000
       metric_buffer_limit = 10000
       collection_jitter = "0s"
       flush_interval = "10s"
       flush_jitter = "0s"
       precision = ""
-      hostname = ""
+      hostname = "$HOSTNAME"
       omit_hostname = false
       debug = false
+      quiet = true
 
     [[outputs.influxdb]]
       urls = ["http://influxdb-svc.monitoring.svc.cluster.local:8086/"]
@@ -62,30 +63,38 @@ data:
       user_agent = "telegraf"
       skip_database_creation = false
 
+    [[inputs.net]]
+    [[inputs.netstat]]
+    [[inputs.diskio]]
+    [[inputs.linux_sysctl_fs]]
+    [[inputs.mem]]
+    [[inputs.processes]]
+    [[inputs.swap]]
+    [[inputs.system]]
+
     [[inputs.cpu]]
       percpu = true
       totalcpu = true
       collect_cpu_time = false
       report_active = false
 
     [[inputs.disk]]
-      ignore_fs = ["tmpfs", "devtmpfs", "devfs", "iso9660", "overlay", "aufs", "squashfs"]
-
-    [[inputs.diskio]]
-
-    [[inputs.kernel]]
-
-    [[inputs.mem]]
-
-    [[inputs.processes]]
-
-    [[inputs.swap]]
-
-    [[inputs.system]]
+      ignore_fs = ["tmpfs",
+        "devtmpfs",
+        "devfs",
+        "iso9660",
+        "overlay",
+        "aufs",
+        "squashfs"
+      ]
 
     [[inputs.dns_query]]
        servers = ["4.2.2.1", "8.8.8.8"]
-       domains = ["www.google.com", "www.twitter.com", "www.amazon.com", "www.wikipedia.org"]
+       domains = ["www.google.com",
+         "www.twitter.com",
+         "www.amazon.com",
+         "www.wikipedia.org"
+       ]
        record_type = "A"
        port = 53
        timeout = 3
@@ -104,12 +113,6 @@ data:
     [[inputs.internal]]
       collect_memstats = true
 
-    [[inputs.linux_sysctl_fs]]
-
-    [[inputs.net]]
-
-    [[inputs.netstat]]
-
     [[inputs.ping]]
       urls = [
           "google.com",
@@ -141,6 +144,19 @@ data:
       timeout = "3m"
       interval = "15m"
       data_format = "csv"
-      csv_column_names = ['Server ID','Sponsor','Server Name','Timestamp','Distance','Ping','Download','Upload','Share','IP Address']
+      csv_column_names = ['Server ID',
+        'Sponsor',
+        'Server Name',
+        'Timestamp',
+        'Distance',
+        'Ping',
+        'Download',
+        'Upload',
+        'Share',
+        'IP Address'
+      ]
       csv_timestamp_column = "Timestamp"
       csv_timestamp_format = "2006-01-02T15:04:05Z07:00"
+
+    [[inputs.docker]]
+      endpoint = "unix:///var/run/docker.sock"