Monitoring framework for minute-by-minute service testing, including live dashboard

This commit is contained in:
m8in
2025-11-27 21:12:30 +01:00
parent 97e8a90b26
commit f383bcec2b
17 changed files with 986 additions and 0 deletions

View File

@@ -0,0 +1,45 @@
#!/bin/bash
_REMOTE_HOST="${1:?"FQDN of server missing: e.g. host.example.net[:port]"}"
_REMOTE_HOSTNAME_FQDN="${_REMOTE_HOST%%:*}" #Removes longest matching pattern ':*' from the end
_REMOTE_HOSTNAME_SHORT="${_REMOTE_HOSTNAME_FQDN%%.*}" #Removes longest matching pattern '.*' from the end
_REMOTE_PORT="${_REMOTE_HOST}:"
_REMOTE_PORT="${_REMOTE_PORT#*:}" #Removes shortest matching pattern '*:' from the begin
_REMOTE_PORT="${_REMOTE_PORT%%:*}" #Removes longest matching pattern ':*' from the end
_REMOTE_PORT="${_REMOTE_PORT:-"22"}"
_REMOTE_USER="monitoring"
_SOCKET='~/.ssh/%r@%h:%p'
function checkOrStartSSHMaster() {
timeout --preserve-status 1 ssh -O check -S ${_SOCKET} -p ${_REMOTE_PORT} ${_REMOTE_USER}@${_REMOTE_HOSTNAME_FQDN} 2>&1 | grep -q -F 'Master running' \
&& return 0
ssh -O stop -S ${_SOCKET} -p ${_REMOTE_PORT} ${_REMOTE_USER}@${_REMOTE_HOSTNAME_FQDN} &> /dev/null
ssh -o ControlMaster=auto \
-o ControlPath=${_SOCKET} \
-o ControlPersist=65 \
-p ${_REMOTE_PORT} \
-f ${_REMOTE_USER}@${_REMOTE_HOSTNAME_FQDN} exit &> /dev/null \
&& return 0
echo "FAIL#SSH connection (setup ok?)"
return 1
}
function testDomain(){
checkOrStartSSHMaster \
|| return 1
local _RESULT="$(ssh -S ${_SOCKET} -p ${_REMOTE_PORT} ${_REMOTE_USER}@${_REMOTE_HOSTNAME_FQDN} 'bash /cis/core/printOwnDomain.sh' 2>&1 1>/dev/null)"
[ -z "${_RESULT}" ] \
&& echo "OK" \
&& return 0
echo "WARNING#Check hosts '/cis/core/printOwnDomain'"
return 0
}
testDomain && exit 0

View File

@@ -0,0 +1,67 @@
#!/bin/bash
_REMOTE_HOST="${1:?"FQDN of server missing: e.g. host.example.net[:port]"}"
_REMOTE_HOSTNAME_FQDN="${_REMOTE_HOST%%:*}" #Removes longest matching pattern ':*' from the end
_REMOTE_HOSTNAME_SHORT="${_REMOTE_HOSTNAME_FQDN%%.*}" #Removes longest matching pattern '.*' from the end
_REMOTE_PORT="${_REMOTE_HOST}:"
_REMOTE_PORT="${_REMOTE_PORT#*:}" #Removes shortest matching pattern '*:' from the begin
_REMOTE_PORT="${_REMOTE_PORT%%:*}" #Removes longest matching pattern ':*' from the end
_REMOTE_PORT="${_REMOTE_PORT:-"22"}"
_REMOTE_USER="monitoring"
_SOCKET='~/.ssh/%r@%h:%p'
function checkOrStartSSHMaster() {
timeout --preserve-status 1 ssh -O check -S ${_SOCKET} -p ${_REMOTE_PORT} ${_REMOTE_USER}@${_REMOTE_HOSTNAME_FQDN} 2>&1 | grep -q -F 'Master running' \
&& return 0
ssh -O stop -S ${_SOCKET} -p ${_REMOTE_PORT} ${_REMOTE_USER}@${_REMOTE_HOSTNAME_FQDN} &> /dev/null
ssh -o ControlMaster=auto \
-o ControlPath=${_SOCKET} \
-o ControlPersist=65 \
-p ${_REMOTE_PORT} \
-f ${_REMOTE_USER}@${_REMOTE_HOSTNAME_FQDN} exit &> /dev/null \
&& return 0
echo "FAIL#SSH connection (setup ok?)"
return 1
}
function checkViaHTTP() {
_STATUS="$(curl -I http://${_REMOTE_HOSTNAME_FQDN} 2>/dev/null | head -n 1 | cut -d$' ' -f2)"
[ "${_STATUS}" == "200" ] \
&& echo "OK" \
&& return 0
return 1
}
function checkViaHTTPS() {
_STATUS="$(curl -k -I https://${_REMOTE_HOSTNAME_FQDN} 2>/dev/null | head -n 1 | cut -d$' ' -f2)"
[ "${_STATUS}" == "200" ] \
&& echo "OK" \
&& return 0
return 1
}
#grep:
# -E Use regexp, '.*' => any chars between 'Active:' and '(running)', the round brackets are escaped.
#cut:
# -d Delimiter, marker where to cut (here ;)
# -f Index of column to show (One based, so there is no -f0)
function checkViaSSH() {
checkOrStartSSHMaster \
|| return 1
_RESULT=$(ssh -S ${_SOCKET} -p ${_REMOTE_PORT} ${_REMOTE_USER}@${_REMOTE_HOSTNAME_FQDN} 'systemctl status nginx.service' | grep -E 'Active:.*\(running\)' | cut -d';' -f2)
! [ -z "${_RESULT}" ] && echo "OK#UPTIME:${_RESULT}" || echo "FAIL"
}
#checkViaHTTP && exit 0
#checkViaHTTPS && exit 0
checkViaSSH && exit 0
exit 1

View File

@@ -0,0 +1,9 @@
#!/bin/bash
_SERVER="${1:?"FQDN of server missing"}"
# -4 Use IPv4
# -W SECONDS Wait seconds for an answer
# -c COUNT_VALUE Count of pings being executed
_RESULT="$(ping -4 -W 1 -c 1 "${_SERVER}" | grep "time=" | cut -d'=' -f4)"
! [ -z "${_RESULT}" ] && echo "OK#RTT: ${_RESULT}" || echo "FAIL#PLEASE USE FALLBACK!"

View File

@@ -0,0 +1,37 @@
#!/bin/bash
function checkPostgresSSLCertificate() {
local _SERVER
_SERVER="${1:?"FQDN of server missing"}"
readonly _SERVER
local _RESULT
_RESULT="$(echo | openssl s_client -starttls postgres -connect "${_SERVER}":5432 -servername "${_SERVER}" 2> /dev/null | openssl x509 -noout -enddate | grep -F 'notAfter=' | cut -d'=' -f2)"
readonly _RESULT
[ -z "${_RESULT}" ] \
&& echo "FAIL#Unable to get cert's end date from ${_SERVER}:5432" \
&& return 1
local _ENDDATE
_ENDDATE="$(date --date="${_RESULT}" --utc +%s)"
readonly _ENDDATE
! echo "${_ENDDATE}" | grep -q -E "^[0-9]*$" \
&& echo "FAIL#Unable to parse end date of certificate" \
&& return 1
local _NOW _REMAINING_DAYS
_NOW="$(date --date now +%s)"
_REMAINING_DAYS="$(( (_ENDDATE - _NOW) / 86400 ))"
readonly _NOW _REMAINING_DAYS
[ -z "${_REMAINING_DAYS}" ] \
&& echo "WARN#Only ${_REMAINING_DAYS} days left" \
&& return 1
echo "OK#${_REMAINING_DAYS} days remaining"
return 0
}
checkPostgresSSLCertificate "${@}" && exit 0 || exit 1

View File

@@ -0,0 +1,62 @@
#!/bin/bash
#curl:
# --connect-timeout SECONDS Maximum time allowed for connection
# -k Allow connections to SSL sites without certs (H)
# -L Follow redirects (H)
# --max-time SECONDS Maximum time allowed for the transfer
# -s Silent mode. Don't output anything
# --head Show head information only
# --no-progress-meter Clean output for grep
#grep:
# -q Quite, no output just status codes
# -F Interpret search term as plain text
function checkUrl() {
local _URL _SEARCH_STRING
_URL="${1:?"URL of site missing"}"
_SEARCH_STRING="${2}"
readonly _URL _SEARCH_STRING
local _RESULT
if [ -z "${_SEARCH_STRING}" ]; then
_RESULT="$(curl --connect-timeout 10 --max-time 10 --no-progress-meter --verbose "${_URL}" 2>&1 | grep -o -E "(expire.*|HTTP.*200 OK)")"
else
_RESULT="$(curl --connect-timeout 10 --max-time 10 --no-progress-meter --verbose "${_URL}" 2>&1 | grep -o -E "(expire.*|HTTP.*200 OK|${_SEARCH_STRING})")"
fi
readonly _RESULT
! echo "${_RESULT}" | grep -q -F '200 OK' \
&& echo "FAIL#Status code 200 not found" \
&& return 1
! [ -z "${_SEARCH_STRING}" ] \
&& ! echo "${_RESULT}" | grep -q -F "${_SEARCH_STRING}" \
&& echo "FAIL#Search string not found" \
&& return 1
local _ENDDATE
_ENDDATE="$(echo "${_RESULT}" | grep -F 'expire' | cut -d':' -f2-)"
_ENDDATE="$(date --date="${_ENDDATE}" --utc +%s)"
readonly _ENDDATE
! echo "${_ENDDATE}" | grep -q -E "^[0-9]*$" \
&& echo "FAIL#Unable to parse end date of certificate" \
&& return 1
local _NOW _REMAINING_DAYS
_NOW="$(date --date now +%s)"
_REMAINING_DAYS="$(( (_ENDDATE - _NOW) / 86400 ))"
readonly _NOW _REMAINING_DAYS
# less than 30 days remaining => should be warned
[ "${_REMAINING_DAYS}" -le "30" ] \
&& echo "WARN#Certificate: only ${_REMAINING_DAYS} days left" \
&& return 1
echo "OK#Certificate: ${_REMAINING_DAYS} days remaining"
return 0
}
#((curl --connect-timeout 10 --max-time 10 -k -s --head --no-progress-meter "${_URL}" | grep -qF '200 OK') && echo OK) || echo FAIL
checkUrl "${1}" "${2}" && exit 0 || exit 1

View File

@@ -0,0 +1,50 @@
#!/bin/bash
_REMOTE_HOST="${1:?"FQDN of server missing: e.g. host.example.net[:port]"}"
_ZFS_POOL="${2:?"Name of zfs pool missing: e.g. zpool1"}"
_REMOTE_HOSTNAME_FQDN="${_REMOTE_HOST%%:*}" #Removes longest matching pattern ':*' from the end
_REMOTE_HOSTNAME_SHORT="${_REMOTE_HOSTNAME_FQDN%%.*}" #Removes longest matching pattern '.*' from the end
_REMOTE_PORT="${_REMOTE_HOST}:"
_REMOTE_PORT="${_REMOTE_PORT#*:}" #Removes shortest matching pattern '*:' from the begin
_REMOTE_PORT="${_REMOTE_PORT%%:*}" #Removes longest matching pattern ':*' from the end
_REMOTE_PORT="${_REMOTE_PORT:-"22"}"
_REMOTE_USER="monitoring"
_SOCKET='~/.ssh/%r@%h:%p'
function checkOrStartSSHMaster() {
timeout --preserve-status 1 ssh -O check -S ${_SOCKET} -p ${_REMOTE_PORT} ${_REMOTE_USER}@${_REMOTE_HOSTNAME_FQDN} 2>&1 | grep -q -F 'Master running' \
&& return 0
ssh -O stop -S ${_SOCKET} -p ${_REMOTE_PORT} ${_REMOTE_USER}@${_REMOTE_HOSTNAME_FQDN} &> /dev/null
ssh -o ControlMaster=auto \
-o ControlPath=${_SOCKET} \
-o ControlPersist=65 \
-p ${_REMOTE_PORT} \
-f ${_REMOTE_USER}@${_REMOTE_HOSTNAME_FQDN} exit &> /dev/null \
&& return 0
echo "FAIL#SSH connection (setup ok?)"
return 1
}
function testPool(){
checkOrStartSSHMaster \
|| return 1
local _RESPONSE="$(ssh -S ${_SOCKET} -p ${_REMOTE_PORT} ${_REMOTE_USER}@${_REMOTE_HOSTNAME_FQDN} 'zpool status ${_ZFS_POOL} | grep -F scrub')"
local _RESULT=$(echo "${_RESPONSE}" | grep -F 'scrub repaired 0B' | grep -F '0 errors')
_RESULT="${_RESULT#*on}" #Removes shortest matching pattern '*on' from the begin
[ -z "${_RESULT}" ] \
&& echo "FAIL#CHECK POOL: ${_ZFS_POOL}" \
&& return 0
echo "OK#Scrubbed on ${_RESULT}."
return 0
}
testPool && exit 0
exit 1

View File

@@ -0,0 +1,106 @@
#!/bin/bash
_SCRIPT="$(readlink -f "${0}" 2> /dev/null)"
# Folders always ends with an tailing '/'
_CIS_ROOT="${_SCRIPT%%/script/monitor/*}/" #Removes longest matching pattern '/script/monitor/*' from the end
_DOMAIN="$("${_CIS_ROOT:?"Missing CIS_ROOT"}core/printOwnDomain.sh")"
_COMPOSITIONS="${_CIS_ROOT:?"Missing CIS_ROOT"}definitions/${_DOMAIN:?"Missing DOMAIN"}/compositions/"
_REMOTE_HOST="${1:?"FQDN of server missing: e.g. host.example.net[:port]"}"
_REMOTE_HOSTNAME_FQDN="${_REMOTE_HOST%%:*}" #Removes longest matching pattern ':*' from the end
_REMOTE_HOSTNAME_SHORT="${_REMOTE_HOSTNAME_FQDN%%.*}" #Removes longest matching pattern '.*' from the end
_REMOTE_PORT="${_REMOTE_HOST}:"
_REMOTE_PORT="${_REMOTE_PORT#*:}" #Removes shortest matching pattern '*:' from the begin
_REMOTE_PORT="${_REMOTE_PORT%%:*}" #Removes longest matching pattern ':*' from the end
_REMOTE_PORT="${_REMOTE_PORT:-"22"}"
_REMOTE_USER="monitoring"
_SOCKET='~/.ssh/%r@%h:%p'
# This is crucial:
# - default value for the filter part is extracted from the first parameter (FQDN)
# - but you can override this part to to adapt the test during a change of the domain.
# (e.g. the short hostname can be an option - or even a better default in the future)
_ZFS_SNAPSHOT_FILTER="@SYNC_${2:-"${_REMOTE_HOSTNAME_FQDN:?"Missing REMOTE_HOSTNAME_FQDN"}"}"
_MODE="${3:-"normal"}"
_NOW_UTC_UNIXTIME=$(date -u +%s)
_DEBUG_PATH="/tmp/monitor/"
function checkOrStartSSHMaster() {
timeout --preserve-status 1 ssh -O check -S ${_SOCKET} -p ${_REMOTE_PORT} ${_REMOTE_USER}@${_REMOTE_HOSTNAME_FQDN} 2>&1 | grep -q -F 'Master running' \
&& return 0
ssh -O stop -S ${_SOCKET} -p ${_REMOTE_PORT} ${_REMOTE_USER}@${_REMOTE_HOSTNAME_FQDN} &> /dev/null
ssh -o ControlMaster=auto \
-o ControlPath=${_SOCKET} \
-o ControlPersist=65 \
-p ${_REMOTE_PORT} \
-f ${_REMOTE_USER}@${_REMOTE_HOSTNAME_FQDN} exit &> /dev/null \
&& return 0
echo "FAIL#SSH connection (setup ok?)"
return 1
}
function checkSync() {
checkOrStartSSHMaster \
|| return 1
[ "${_MODE}" == "debug" ] \
&& mkdir -p "${_DEBUG_PATH}" > /dev/null \
&& echo "Now: ${_NOW_UTC_UNIXTIME}" > ${_DEBUG_PATH}SECONDS_BEHIND_${_REMOTE_HOSTNAME_FQDN}.txt
! [ -d "${_COMPOSITIONS:?"Missing COMPOSITIONS"}" ] \
&& echo "WARN#no compositions" \
&& return 0
[ "${_MODE}" == "debug" ] \
&& echo "Snapshot filter: ${_ZFS_SNAPSHOT_FILTER}" >> ${_DEBUG_PATH}SECONDS_BEHIND_${_REMOTE_HOSTNAME_FQDN}.txt
# This retrieves the list of the interesting snapshots including creation timestamp
_SNAPSHOTS="$(ssh -S ${_SOCKET} -p ${_REMOTE_PORT} ${_REMOTE_USER}@${_REMOTE_HOSTNAME_FQDN} zfs list -po creation,name -r -t snapshot zpool1/persistent | grep -F ${_ZFS_SNAPSHOT_FILTER})"
[ "${_MODE}" == "debug" ] \
&& echo "${_SNAPSHOTS}" > ${_DEBUG_PATH}SNAPSHOTS_${_REMOTE_HOSTNAME_FQDN}.txt
[ -z "${_SNAPSHOTS}" ] \
&& echo "FAIL#no snapshots" \
&& return 1
echo "OK#Checks running"
for _COMPOSITION_PATH in ${_COMPOSITIONS}*; do
# If remote host is found than it is responsible for this container-composition, otherwise skip
# (grep -E "^[[:blank:]]*something" means. Line has to start with "something", leading blank chars are ok.)
grep -E "^[[:blank:]]*${_REMOTE_HOSTNAME_SHORT}" "${_COMPOSITION_PATH}/zfssync-hosts" &> /dev/null \
|| continue;
_COMPOSITION_NAME="${_COMPOSITION_PATH##*/}" #Removes longest matching pattern '*/' from the begin
_LAST_SNAPSHOT_UNIXTIME="$(echo "${_SNAPSHOTS}" | grep ${_COMPOSITION_NAME} | tail -n 1 | cut -d' ' -f1)"
_SECONDS_BEHIND=$[ ${_NOW_UTC_UNIXTIME} - ${_LAST_SNAPSHOT_UNIXTIME} ]
[ "${_MODE}" == "debug" ] \
&& echo "${_LAST_SNAPSHOT_UNIXTIME} ${_COMPOSITION_NAME} on ${_REMOTE_HOSTNAME_FQDN} behind: ${_SECONDS_BEHIND}s" >> ${_DEBUG_PATH}SECONDS_BEHIND_${_REMOTE_HOSTNAME_FQDN}.txt
[ "${_SECONDS_BEHIND}" -lt 40 ] \
&& continue
[ "${_SECONDS_BEHIND}" -lt 60 ] \
&& echo "ZFSSYNC_of_${_REMOTE_HOSTNAME_SHORT}_LAGGING?WARN#${_COMPOSITION_NAME} ${_SECONDS_BEHIND}s" \
&& continue
echo "ZFSSYNC_of_${_REMOTE_HOSTNAME_SHORT}_LAGGING?FAIL#${_COMPOSITION_NAME} ${_SECONDS_BEHIND}s"
done
}
RESULTS="$(checkSync)"
[ "${_MODE}" == "debug" ] \
&& echo "$RESULTS" > ${_DEBUG_PATH}RESULTS_${_REMOTE_HOSTNAME_FQDN}.txt
echo "$RESULTS"

View File

@@ -0,0 +1,57 @@
#!/bin/bash
_REMOTE_HOST="${1:?"FQDN of server missing: e.g. host.example.net[:port]"}"
_REMOTE_HOSTNAME_FQDN="${_REMOTE_HOST%%:*}" #Removes longest matching pattern ':*' from the end
_REMOTE_HOSTNAME_SHORT="${_REMOTE_HOSTNAME_FQDN%%.*}" #Removes longest matching pattern '.*' from the end
_REMOTE_PORT="${_REMOTE_HOST}:"
_REMOTE_PORT="${_REMOTE_PORT#*:}" #Removes shortest matching pattern '*:' from the begin
_REMOTE_PORT="${_REMOTE_PORT%%:*}" #Removes longest matching pattern ':*' from the end
_REMOTE_PORT="${_REMOTE_PORT:-"22"}"
_REMOTE_USER="monitoring"
_SOCKET='~/.ssh/%r@%h:%p'
function checkOrStartSSHMaster() {
timeout --preserve-status 1 ssh -O check -S ${_SOCKET} -p ${_REMOTE_PORT} ${_REMOTE_USER}@${_REMOTE_HOSTNAME_FQDN} 2>&1 | grep -q -F 'Master running' \
&& return 0
ssh -O stop -S ${_SOCKET} -p ${_REMOTE_PORT} ${_REMOTE_USER}@${_REMOTE_HOSTNAME_FQDN} &> /dev/null
ssh -o ControlMaster=auto \
-o ControlPath=${_SOCKET} \
-o ControlPersist=65 \
-p ${_REMOTE_PORT} \
-f ${_REMOTE_USER}@${_REMOTE_HOSTNAME_FQDN} exit &> /dev/null \
&& return 0
echo "FAIL#SSH connection (setup ok?)"
return 1
}
function testSpace(){
checkOrStartSSHMaster \
|| return 1
local _RESULT="$(ssh -S ${_SOCKET} -p ${_REMOTE_PORT} ${_REMOTE_USER}@${_REMOTE_HOSTNAME_FQDN} 'zpool list -H -o capacity,name')"
local _SPACE_USED=$(echo "${_RESULT}" | /usr/bin/tail -n 1 | /usr/bin/cut -f1)
local _POOL=$(echo "${_RESULT}" | /usr/bin/tail -n 1 | /usr/bin/cut -f2)
[ -z "${_SPACE_USED}" ] \
&& echo "FAIL#NO value" \
&& return 0
[ "${1:?"Missing OK_THRESHOLD"}" -ge "${_SPACE_USED%\%*}" ] \
&& echo "OK#${_SPACE_USED} used ${_POOL}." \
&& return 0
[ "${2:?"Missing INFO_THRESHOLD"}" -ge "${_SPACE_USED%\%*}" ] \
&& echo "INFO#${_SPACE_USED} already used ${_POOL}." \
&& return 0
echo "FAIL#${_SPACE_USED} used ${_POOL}!"
return 0
}
testSpace 80 90 && exit 0
exit 1