AWS RDS disk usage check for Naemon/Nagios

check_rds_disk.sh

#!/bin/bash
# size units are in GB
# syntax: check_rds_disk.sh <1:db ident> <2:max delay> <3:warn> <4:crit>
#    e.g: check_rds_disk.sh rds.db1 300 70 85

db_name="${1/*./}"
# max_daly in seconds
max_delay=$2
warn=$3
crit=$4

base_delay=360

msg_append(){
        if [ -z "$msg" ]; then
                msg="$1"
        else
                msg="$msg, $1"
        fi
}

ec_esca(){ [ ${ec:-0} -lt $1 ] && ec=$1; }

pptime(){
        if [ $1 -gt 129600 ]; then
                echo "$(($1/86400))d$(($1%86400/3600))h$(($1%3600/60))m"
        elif [ $1 -gt 5400 ]; then
                echo "$(($1/3600))h$(($1%3600/60))m"
        elif [ $1 -gt 120 ]; then
                echo "$(($1/60))m$(($1%60))s"
        else
                echo "$(($1))s"
        fi
}

get_store(){ aws rds describe-db-instances --db-instance-identifier $db_name | jq '.DBInstances[].AllocatedStorage'; }

get_cw(){
        unset o
        # retries for cloudwatch eventual consistancy (in a monitoring service?!?!?!?)
        for i in `seq 1 10`; do
                o=$(aws cloudwatch get-metric-statistics --namespace AWS/RDS \
                        --metric-name FreeStorageSpace \
                        --dimensions Name=DBInstanceIdentifier,Value=$db_name \
                        --start-time $((`date +%s`/60*60-60*$i)) --end-time $((`date +%s`/60*60-60*($i-1))) \
                        --period 60 --statistics Average --output text | sed 1d |\
                        awk '{printf "%s %d %s %s", $1, $2/1024/1024/1024, $3, $4}')
                [ ! -z "$o" ] && break
        done
        if [ ! -z "$o" ]; then
                read D m t u <<< "$o"
                echo ${m/.*/} $((`date +%s`-`date -d "$t" +%s`))
        else
                echo "NA NA"
        fi
}

size=`get_store`
read data delay <<< "`get_cw`"
perc=$((($size-$data)*100/$size))

perf_data="freeGB:$data; delaySec:$delay; sizeGB:$size; usedPercent:$perc"

### check delay
###############

if [ ${delay} -gt $max_delay ]; then
        msg_append "delay too big (`pptime $delay`>`pptime $max_delay`)"
        ec_esca 2
fi

### check data
#######################

if [ "$data" = "NA" ]; then
        msg_append "unable to get cloudwatch data point"
        ec_esca 2
elif [ $perc -gt $crit ]; then
        msg_append "Used storage ${perc}% > ${crit}%"
        ec_esca 2
elif [ $perc -gt $warn ]; then
        msg_append "Used storage ${perc}% > ${warn}%"
        ec_esca 1
else
        msg_append "Used storage ${perc}% < ${warn}%"
        ec_esca 0
fi

echo "$msg | $perf_data"
#echo "<pre>$info</pre>"
exit $ec