prometheus/node_exporter · error

failed to retrieve rapl stats

Error message

failed to retrieve rapl stats: %w

What it means

The RAPL power collector wraps any non-permission error from reading powercap zones (procfs powercap API) with 'failed to retrieve rapl stats'. Missing powercap and permission errors are downgraded to ErrNoData or a debug log, so this error means a genuine failure enumerating or reading the Intel RAPL powercap interface outside those tolerated cases.

Solutions

1. Inspect /sys/class/powercap for intact intel-rapl:* entries with readable name/energy_uj files
2. If RAPL is unusable on this host, disable the rapl collector instead of scraping it
3. Check dmesg for intel_rapl/powercap errors and update the kernel or run on bare metal
4. Verify the container runtime is not masking parts of /sys/class/powercap

Example fix

// before (host without usable RAPL)
node_exporter --collector.rapl
// after
node_exporter --collector.rapl=false  # or omit --collector.rapl when not using --collector.disable-defaults

// caller-side pre-check before enabling the collector
if entries, err := os.ReadDir("/sys/class/powercap"); err != nil || len(entries) == 0 {
	// RAPL unavailable; do not enable --collector.rapl
}

func raplAvailable() bool {
	_, err := os.Stat("/sys/class/powercap/intel-rapl:0")
	return err == nil
}

if err := collector.Update(ch); err != nil {
	if strings.Contains(err.Error(), "failed to retrieve rapl stats") {
		logger.Warn("RAPL unavailable on this host; power metrics skipped", "err", err)
	} else {
		return err
	}
}

When it happens

Trigger: powercap zone initialization returning an error other than os.ErrNotExist/os.ErrPermission during Update, e.g. malformed /sys/class/powercap contents or I/O errors reading energy files.

Common situations: Containers with partial /sys/class/powercap visibility; virtualized hosts exposing a broken powercap tree; kernels with partial or faulty intel_rapl support.

Understand the failure class

Background: "failed to read file", EACCES, ENOENT and "could not read <path>" errors: when a program can't read a file from disk — this error's family across 49 libraries.

Related errors

• Could not derive a human-readable chip type for
• failed to open sysfs
• failed to retrieve bcache stats
• failed to open sysfs
• failed to retrieve bcachefs stats

AI-assisted analysis of prometheus/node_exporter@17ddd77c59 (2026-09-07). Data as JSON: /api/errors/6b5dc922aa97fca5. Report an issue: GitHub.

Appendix: source

Thrown at collector/rapl_linux.go:82

		joulesMetricDesc: joulesMetricDesc,
	}
	return &collector, nil
}

// Update implements Collector and exposes RAPL related metrics.
func (c *raplCollector) Update(ch chan<- prometheus.Metric) error {
	// nil zones are fine when platform doesn't have powercap files present.
	zones, err := sysfs.GetRaplZones(c.fs)
	if err != nil {
		if errors.Is(err, os.ErrNotExist) {
			c.logger.Debug("Platform doesn't have powercap files present", "err", err)
			return ErrNoData
		}
		if errors.Is(err, os.ErrPermission) {
			c.logger.Debug("Can't access powercap files", "err", err)
			return ErrNoData
		}
		return fmt.Errorf("failed to retrieve rapl stats: %w", err)
	}

	for _, rz := range zones {
		microJoules, err := rz.GetEnergyMicrojoules()
		if err != nil {
			if errors.Is(err, os.ErrPermission) {
				c.logger.Debug("Can't access energy_uj file", "zone", rz, "err", err)
				return ErrNoData
			}
			return err
		}

		joules := float64(microJoules) / 1000000.0

		if *raplZoneLabel {
			ch <- c.joulesMetricWithZoneLabel(rz, joules)
		} else {
			ch <- c.joulesMetric(rz, joules)

View on GitHub (pinned to 17ddd77c59)