Add replication monitoring

Author: Federico Ceratto

ansible/roles/node_exporter/files/db_replication_mon_active

@@ -0,0 +1,22 @@
+#!/bin/bash
+# Deployed by ansible
+# See roles/node_exporter/files/db_replication_mon_active
+# active database --> standby
+#
+# Errors are reported as RTT=0
+
+set -ux
+FN=/run/nodeexp/db_replication_socket.prom
+while true; do
+  socket_rtt=$(ss -ntpi state established 'dst 37.218.242.175' | tr -s " " "\n" | grep ^rtt: | cut -c5- | cut -d'/' -f1)
+  socket_rtt=${socket_rtt:-0}
+  ping_rtt=$(ping -w 1 -c 1 10.1.0.1 | tail -n1  | cut -d'/' -f5)
+  ping_rtt=${ping_rtt:-0}
+
+  cat <<EOF > $FN.$$
+db_replication_rtt{role="standby"} $socket_rtt
+ping_rtt{role="standby"} $ping_rtt
+EOF
+  mv $FN.$$ $FN
+  sleep 1
+done

ansible/roles/node_exporter/files/db_replication_mon_standby

@@ -0,0 +1,22 @@
+#!/bin/bash
+# Deployed by ansible
+# See roles/node_exporter/files/db_replication_mon_standby
+# standby database --> active
+#
+# Errors are reported as RTT=0
+
+set -ux
+FN=/run/nodeexp/db_replication_socket.prom
+while true; do
+  socket_rtt=$(ss -ntpi state established 'dst 37.218.240.56 dport = 5432' | tr -s " " "\n" | grep ^rtt: | cut -c5- | cut -d'/' -f1)
+  socket_rtt=${socket_rtt:-0}
+  ping_rtt=$(ping -w 1 -c 1 10.1.0.2 | tail -n1  | cut -d'/' -f5)
+  ping_rtt=${ping_rtt:-0}
+
+  cat <<EOF > $FN.$$
+db_replication_rtt{role="standby"} $socket_rtt
+ping_rtt{role="standby"} $ping_rtt
+EOF
+  mv $FN.$$ $FN
+  sleep 1
+done

ansible/roles/node_exporter/handlers/main.yml

@@ -3,14 +3,22 @@
   service:
     name: node_exporter
     state: restarted
+
 - name: restart seeksample
   service:
     name: seeksample
     state: restarted
+
 - name: restart tcpmetrics
   service:
     name: tcpmetrics
     state: restarted
+
+- name: restart db_replication_mon
+  service:
+    name: db_replication_mon
+    state: restarted
+
 - name: restart collector-metrics
   service:
     name: collector-metrics

ansible/roles/node_exporter/tasks/main.yml

@@ -89,6 +89,26 @@
     dest: /etc/systemd/system/tcpmetrics.service
   when: "'have_tcpmetrics' in group_names"
 
+- name: Install db_replication_mon on active DB
+  copy: src=db_replication_mon_active dest=/srv/db_replication_mon mode=0555
+  when: "'db_active' in group_names"
+  tags: db_replication_mon
+
+- name: Install db_replication_mon on standby DB
+  copy: src=db_replication_mon_standby dest=/srv/db_replication_mon mode=0555
+  when: "'db_standby' in group_names"
+  tags: db_replication_mon
+
+- name: Install db_replication_mon systemd service file
+  notify:
+  - restart db_replication_mon
+  - systemctl daemon-reload
+  template:
+    src: db_replication_mon.service
+    dest: /etc/systemd/system/db_replication_mon.service
+  when: "'db_active' in group_names or 'db_active' in group_names"
+  tags: db_replication_mon
+
 - name: Install collector-metrics systemd service file
   notify:
   - restart collector-metrics

ansible/roles/node_exporter/templates/db_replication_mon.service

@@ -0,0 +1,27 @@
+# Deployed by ansible
+# see roles/node_exporter/templates/db_replication_mon.service
+[Unit]
+Description=DB replication monitor
+
+[Service]
+User=root
+Group=root
+
+ExecStartPre=/usr/bin/install --owner root --group root --mode 0755 --directory {{ node_exporter_textfiles_path }}
+ExecStart=/srv/db_replication_mon
+KillMode=process
+Restart=always
+
+# Hardening
+CapabilityBoundingSet=CAP_SETUID CAP_SETGID
+SystemCallFilter=~@clock @debug @cpu-emulation @keyring @module @mount @obsolete @raw-io @reboot @swap
+NoNewPrivileges=yes
+PrivateDevices=yes
+PrivateTmp=yes
+ProtectHome=yes
+ProtectSystem=full
+ProtectKernelModules=yes
+ProtectKernelTunables=yes
+
+[Install]
+WantedBy=multi-user.target

ansible/roles/plpsql/README.adoc

@@ -0,0 +1,6 @@
+PostgreSQL metadb database
+
+Deploys a temporary traffic forwarder using A VPN
+  HKG -> mia-ps-test.ooni.nu -> AMS
+
+Runs db_replication_mon.service on HKG and AMS db to monitor the replication