Problem: When the node_exporter_metrics input plugin fails to read hardware sensors (hwmon, thermal zones, cpufreq), the error logs only show:
[error] [ne_utils.c:132 errno=61] No data available
[error] [ne_utils.c:132 errno=6] No such device or address
This is useless for debugging because users have no way to know which file/path is missing.
Expected: Error messages should include the file path, e.g.:
[error] [input:node_exporter_metrics:node_exporter_metrics.0] could not open '/sys/class/hwmon/hwmon4/fan1_label'
[error] [input:node_exporter_metrics:node_exporter_metrics.0] could not read from '/sys/class/thermal/thermal_zone6/temp'
This is particularly problematic on:
- Virtual machines (missing cpufreq, hwmon sensors)
- Laptops with varying hardware configurations
- Systems with optional thermal zones
Environment: Linux, Fluent Bit v5.0.7
Additional context: The issue was identified when debugging missing hwmon/cpufreq/thermal sensors on a laptop with many hwmon devices (some optional).
Problem: When the node_exporter_metrics input plugin fails to read hardware sensors (hwmon, thermal zones, cpufreq), the error logs only show:
This is useless for debugging because users have no way to know which file/path is missing.
Expected: Error messages should include the file path, e.g.:
This is particularly problematic on:
Environment: Linux, Fluent Bit v5.0.7
Additional context: The issue was identified when debugging missing hwmon/cpufreq/thermal sensors on a laptop with many hwmon devices (some optional).