vectordotdev/vector · error

bins is empty

Error message

bins is empty

What it means

find_quantile implements R-3 quantile estimation over a cumulative histogram of samples. It calls bins.last().expect("bins is empty") to read the total count from the final cumulative bin; if the sample slice is empty there is no last element and the expect panics. The function assumes at least one Sample was collected before any quantile is queried.

Solutions

1. Guard in from_samples: return early (f64::NAN or None) when samples is empty before calling find_quantile.
2. Ensure the histogram is only finalized when it contains at least one sample.
3. Fix upstream filters/windowing so the aggregation window is not empty when statistics are requested.

Example fix

// before
let count = bins.last().expect("bins is empty").rate;
// after
let Some(last) = bins.last() else { return f64::NAN; };
let count = last.rate;

if samples.is_empty() { /* skip quantile computation */ }

fn has_samples(bins: &[Sample]) -> bool { !bins.is_empty() }

When it happens

Trigger: Calling find_quantile (via from_samples) with an empty &[Sample] slice — i.e. a histogram built with zero observations, typically when an aggregation window produced no data.

Common situations: Computing p50/p95/p99 statistics in the aggregator over a time window where no events arrived (idle source, misconfigured metric tags, or a transform filtering everything out).

Understand the failure class

Background: "must not be empty", "cannot be empty" — required-field validation errors across open-source libraries — this error's family across 41 libraries.

Related errors

• a record with a next ID must have an event count
• a valid HTTP/1 URI is valid as an HTTP URI
• a validated HTTP endpoint is a valid `http 1` URI
• all json-file keys should be matched
• allowed_lateness_ms validated to fit in i64 in…

AI-assisted analysis of vectordotdev/vector@bdb87aeaa4 (2026-09-16). Data as JSON: /api/errors/e8b5149f0e03b76f. Report an issue: GitHub.

Appendix: source

Thrown at src/sinks/util/statistic.rs:90

                    median,
                    avg,
                    sum,
                    count: count as u64,
                    quantiles,
                }
            }),
        }
    }
}

/// `bins` is a cumulative histogram
/// We are using R-3 (without choosing the even integer in the case of a tie),
/// it might be preferable to use a more common function, such as R-7.
///
/// List of quantile functions:
/// <https://en.wikipedia.org/wiki/Quantile#Estimating_quantiles_from_a_sample>
fn find_quantile(bins: &[Sample], p: f64) -> f64 {
    let count = bins.last().expect("bins is empty").rate;
    find_sample(bins, (p * count as f64).round() as u32)
}

/// `bins` is a cumulative histogram
/// Return the i-th smallest value,
/// i starts from 1 (i == 1 mean the smallest value).
/// i == 0 is equivalent to i == 1.
fn find_sample(bins: &[Sample], i: u32) -> f64 {
    let index = match bins.binary_search_by_key(&i, |sample| sample.rate) {
        Ok(index) => index,
        Err(index) => index,
    };
    bins[index].value
}

pub fn validate_quantiles(quantiles: &[f64]) -> Result<(), ValidationError> {
    if quantiles
        .iter()

View on GitHub (pinned to bdb87aeaa4)