Code Quality: Metric parsing silently uses default values

[drmingdrmer]

Summary

Metric value parsing silently falls back to 0.0 and histogram bucket parsing silently drops invalid entries.

Locations

crates/service/src/meta_node/meta_node.rs:509

let value: f64 = value_str.parse().unwrap_or(0.0);

crates/service/src/meta_node/meta_node.rs:714-718

.filter_map(|(le, cumulative_count)| {
    le.parse::<f64>()
        .ok()
        .map(|le_val| (le_val, *cumulative_count))
})

Problem

• Invalid metric values become 0.0 silently
• Invalid histogram bucket labels are dropped silently
• Data corruption or format issues are masked
• Monitoring dashboards may show incorrect data

Impact

• Metrics dashboards show incorrect values
• Capacity planning based on wrong data
• Format changes go unnoticed
• Data corruption is hidden

Suggested Fix

Log invalid values:

let value: f64 = value_str.parse().unwrap_or_else(|e| {
    warn!("Invalid metric value '{}': {}", value_str, e);
    0.0
});

// For histogram buckets
.filter_map(|(le, cumulative_count)| {
    match le.parse::<f64>() {
        Ok(le_val) => Some((le_val, *cumulative_count)),
        Err(e) => {
            warn!("Invalid histogram bucket '{}': {}", le, e);
            None
        }
    }
})

Priority

P3 - Observability improvement