Skip to content
Sarj Code Standards

failed-probe-zero-threshold

Do not let a failed subprocess count probe meet its threshold through a zero fallback.

SARJ538 Rule source Tests Catalog JSON

Why

A failed observation is different from an observed count of zero. Coalescing the error sentinel can make a recovery comparison succeed when its baseline is zero.

Fix

Check the probe result for None before comparing it, or make an intentional zero-baseline skip policy explicit.

Examples

Before — flagged Keep failed observation distinct from zero
app/recovery.py
import subprocess




def read_status():
    try:
        result = subprocess.run(
            ["service", "status"], capture_output=True, text=True, check=False
        )
    except (subprocess.TimeoutExpired, OSError):
        return None
    return None if result.returncode else result.stdout




def active_count():
    output = read_status()
    return None if output is None else output.count("active")




def recovered(baseline: int) -> bool:
    if (active_count() or 0) >= baseline:
        return True
    return False
After — preferred Reject a missing observation explicitly
app/recovery.py
import subprocess




def read_status():
    try:
        result = subprocess.run(
            ["service", "status"], capture_output=True, text=True, check=False
        )
    except (subprocess.TimeoutExpired, OSError):
        return None
    return None if result.returncode else result.stdout




def active_count():
    output = read_status()
    return None if output is None else output.count("active")




def recovered(baseline: int) -> bool:
    observed = active_count()
    if observed is not None and observed >= baseline:
        return True
    return False