Scripting and CI

witr is built to be scripted, not only read. Two things make that work: --json and real exit codes.

JSON output#

witr --port 8000 --json

Works with every target type and with multiple mixed targets in one call. Combine with jq to pull out exactly the field a check needs.

Branching on the exit code#

witr nginx --short
case $? in
  0) echo "All clear" ;;
  1) echo "Warnings detected" ;;
  2) echo "Process not running" ;;
  3) echo "Need elevated privileges" ;;
  4) echo "Invalid input or ambiguous match" ;;
  5) echo "Internal error" ;;
esac

The distinction between 2 (not running) and 3 (cannot tell, no permission) matters in a monitoring context, where treating a permission failure as an outage produces false alarms. See Exit Codes for the full table.

A deploy pre-flight check#

if witr --port 8000 --short > /dev/null 2>&1; then
  echo "Port 8000 is occupied. Chain:"
  witr --port 8000 --short
  exit 1
fi

The chain in the failure message means whoever reads the CI log gets the cause, not just the symptom.

Watching for warnings#

witr --pid "$APP_PID" --warnings --json

Exit code 1 means at least one warning fired — a public bind, a restart loop, a deleted binary after an upgrade that never restarted the service.

Output modes that combine#

Every mode works with multiple inputs and with each other where it makes sense:

Full list in the flags reference.