check_dns: better time measurement

alpine does not output ms using date, therefore we use perl to get a more
accurate measurement of the dns response time.
The script output is now even more similar to nagios check_dns.
This commit is contained in:
maxi322 2025-08-29 08:14:32 +02:00
parent 0b0a65a3f3
commit 1469953b52

View file

@ -19,19 +19,19 @@ if [ -z "$HOST" ]; then
fi
# run dig and measure the time it takes to run
START_TIME=$(date +%s%3N)
START_TIME=$(perl -MTime::HiRes -e 'print Time::HiRes::time')
dig_output=$(dig +short +timeout=2 +tries=1 "$HOST" @"$SERVER" 2>/dev/null)
dig_rc=$?
END_TIME=$(perl -MTime::HiRes -e 'print Time::HiRes::time')
dig_output_ips=$(echo "$dig_output" | grep -E '^[0-9.]+$' | sort | paste -sd ',' -)
END_TIME=$(date +%s%3N)
ELAPSED_TIME=$((END_TIME - START_TIME))
ELAPSED_TIME=$(perl -e "printf('%.3f', $END_TIME - $START_TIME)")
# validate and perform nagios like output and exit codes
if [ $dig_rc -ne 0 ] || [ -z "$dig_output" ]; then
echo "Domain $HOST was not found by the server"
exit 2
elif [ $dig_rc -eq 0 ]; then
echo "DNS OK: $ELAPSED_TIME ms response time. $HOST returns $dig_output_ips"
echo "DNS OK: $ELAPSED_TIME seconds response time. $HOST returns $dig_output_ips"
exit 0
else
echo "Unknown error"