1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133

use std::slice::Iter;
use std::cmp::Ordering::Equal;

pub type Seconds = u8;

pub use self::Metric::*;

/// Metric collected from a collector to be recorded in a store.
#[derive(Clone, Debug, PartialEq)]
pub enum Metric {
    Count(Dimension, u64),
    Measure(Dimension, f64),
    Sample(Dimension, f64),
}

/// Metrics can grouped by multiple values. Right now that limited to just
/// their name.
#[derive(Clone, Debug, Eq, PartialEq, Hash)]
pub struct Dimension {
    pub name: String,
    pub source: Option<String>,
}

impl Dimension {
    pub fn with_name<S: AsRef<str>>(name: S) -> Dimension {
        Dimension {
            name: name.as_ref().to_owned(),
            source: None,
        }
    }

    pub fn with_name_and_source<N: AsRef<str>, S: AsRef<str>>(name: N, source: S) -> Dimension {
        Dimension {
            name: name.as_ref().to_owned(),
            source: Some(source.as_ref().to_owned()),
        }
    }

    /// Returns a new `Dimension` with the given name; all other fields will
    /// be copied over.
    pub fn renamed<S: AsRef<str>>(&self, name: S) -> Dimension {
        Dimension {
            name: name.as_ref().to_owned(),
            source: self.source.clone(),
        }
    }
}

#[derive(Debug, PartialEq)]
pub enum AggregatedMetricType {
    Count,
    Measure,
    Sample,
}

/// The final value resulting from aggregating a metric's values.
pub type AggregatedMetric = (AggregatedMetricType, Dimension, f64);

/// All the metrics in a given time interval coalesced into a single value for
/// each metric.
#[derive(Debug, PartialEq)]
pub struct AggregatedMetrics {
    metrics: Vec<AggregatedMetric>,
}

impl AggregatedMetrics {
    pub fn new() -> AggregatedMetrics {
        AggregatedMetrics {
            metrics: vec![],
        }
    }

    pub fn with_metrics(metrics: Vec<AggregatedMetric>) -> AggregatedMetrics {
        AggregatedMetrics {
            metrics: metrics,
        }
    }

    pub fn aggregate_counts<'a, I>(&mut self, counts: I)
        where I: Iterator<Item=(&'a Dimension, &'a u64)>
    {
        for (dim, value) in counts {
            self.metrics.push((AggregatedMetricType::Count, dim.to_owned(), *value as f64))
        }
    }

    /// Rolls up all the given measures. The minimum, maximum, median,
    /// average (mean), and 95th percentile summary measures will all be
    /// emitted, as well as a total count of all the individual measures
    /// received in the period.
    pub fn aggregate_measures<'a, I>(&mut self, measures: I)
        where I: Iterator<Item=(&'a Dimension, &'a Vec<f64>)>
    {
        use self::AggregatedMetricType::*;

        for (dim, values) in measures {
            let mut sorted = values.clone();
            sorted.sort_by(|a, b| a.partial_cmp(b).unwrap_or(Equal));

            let min          = *sorted.first().unwrap();
            let max          = *sorted.last().unwrap();
            let median       = sorted[sorted.len() / 2]; // TODO: Improve how we calculate the median
            let average      = sorted.iter().fold(0.0, |sum, val| { sum + val }) / (sorted.len() as f64);
            let percentile95 = sorted[(sorted.len() as f64 * 0.95) as usize];
            let percentile99 = sorted[(sorted.len() as f64 * 0.99) as usize];

            self.metrics.push((Measure, dim.renamed(format!("{}.min",          dim.name)), min));
            self.metrics.push((Measure, dim.renamed(format!("{}.max",          dim.name)), max));
            self.metrics.push((Measure, dim.renamed(format!("{}.median",       dim.name)), median));
            self.metrics.push((Measure, dim.renamed(format!("{}.avg",          dim.name)), average));
            self.metrics.push((Measure, dim.renamed(format!("{}.95percentile", dim.name)), percentile95));
            self.metrics.push((Measure, dim.renamed(format!("{}.99percentile", dim.name)), percentile99));

            self.metrics.push((Count,   dim.renamed(format!("{}.count", dim.name)), sorted.len() as f64));
        }
    }

    pub fn aggregate_samples<'a, I>(&mut self, samples: I)
        where I: Iterator<Item=(&'a Dimension, &'a f64)>
    {
        for (dim, value) in samples {
            self.metrics.push((AggregatedMetricType::Sample, dim.to_owned(), *value as f64))
        }
    }

    pub fn iter(&self) -> Iter<AggregatedMetric> {
        self.metrics.iter()
    }

    pub fn len(&self) -> usize {
        self.metrics.len()
    }
}