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
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
// Copyright (c) The Libra Core Contributors
// SPDX-License-Identifier: Apache-2.0

//! # Metrics
//!
//! ## Counters
//!
//! Used to measure values that are added to over time, rates
//! can then be used to check how quickly it changes in graphs.
//! An example would be to add every time an incoming message occurs.
//! ```
//! use prometheus::register_int_counter_vec;
//!
//! register_int_counter_vec!(
//!     "name",
//!     "description",
//!     &["dimension_1", "dimension_2"]
//! );
//! ```
//!
//! ## Gauges
//! Used to measure values that change level over time.  An example
//! would be to set the number of connected peers.
//! ```
//! use prometheus::register_int_gauge_vec;
//!
//! register_int_gauge_vec!(
//!     "name",
//!     "description",
//!     &["dimension_1", "dimension_2"]
//! );
//! ```
//!
//! ## Histograms
//! Used to measure histogram values.  An example is network
//! connection latency.
//! ```
//! use prometheus::register_histogram_vec;
//!
//! register_histogram_vec!(
//!     "name",
//!     "description",
//!     &["dimension_1", "dimension_2"]
//! );
//! ```

#![forbid(unsafe_code)]
#![recursion_limit = "128"]

mod json_encoder;
mod json_metrics;
pub mod metric_server;
mod public_metrics;

mod op_counters;
pub use op_counters::{DurationHistogram, OpMetrics};

#[cfg(test)]
mod unit_tests;

// Re-export counter types from prometheus crate
pub use libra_metrics_core::{
    register_histogram, register_histogram_vec, register_int_counter, register_int_counter_vec,
    register_int_gauge, register_int_gauge_vec, Histogram, HistogramTimer, HistogramVec,
    IntCounter, IntCounterVec, IntGauge, IntGaugeVec,
};

use anyhow::Result;
use libra_logger::prelude::*;
use once_cell::sync::Lazy;
use prometheus::{proto::MetricType, Encoder, TextEncoder};
use std::{
    collections::HashMap,
    fs::{create_dir_all, File, OpenOptions},
    io::Write,
    path::Path,
    thread, time,
};

pub static NUM_METRICS: Lazy<IntCounterVec> = Lazy::new(|| {
    register_int_counter_vec!(
        "libra_metrics",
        "Number of metrics in certain states",
        &["type"]
    )
    .unwrap()
});

fn get_metrics_file<P: AsRef<Path>>(dir_path: &P, file_name: &str) -> File {
    create_dir_all(dir_path).expect("Create metrics dir failed");

    let metrics_file_path = dir_path.as_ref().join(file_name);

    info!("Using metrics file {}", metrics_file_path.display());

    OpenOptions::new()
        .append(true)
        .create(true)
        .open(metrics_file_path)
        .expect("Open metrics file failed")
}

pub fn gather_metrics() -> Vec<prometheus::proto::MetricFamily> {
    let metric_families = libra_metrics_core::gather();
    let mut total: u64 = 0;
    let mut families_over_1000: u64 = 0;

    // Take metrics of metric gathering so we know possible overhead of this process
    for metric_family in &metric_families {
        let family_count = metric_family.get_metric().len();
        if family_count > 1000 {
            families_over_1000 = families_over_1000.saturating_add(1);
            let name = metric_family.get_name();
            warn!(
                count = family_count,
                metric_family = name,
                "Metric Family '{}' over 1000 dimensions '{}'",
                name,
                family_count
            );
        }
        total = total.saturating_add(family_count as u64);
    }

    // These metrics will be reported on the next pull, rather than create a new family
    NUM_METRICS.with_label_values(&["total"]).inc_by(total);
    NUM_METRICS
        .with_label_values(&["families_over_1000"])
        .inc_by(families_over_1000);

    metric_families
}

fn get_all_metrics_as_serialized_string() -> Result<Vec<u8>> {
    let all_metrics = gather_metrics();

    let encoder = TextEncoder::new();
    let mut buffer = Vec::new();
    encoder.encode(&all_metrics, &mut buffer)?;
    Ok(buffer)
}

pub fn get_all_metrics() -> HashMap<String, String> {
    // TODO: use an existing metric encoder (same as used by
    // prometheus/metric-server)
    let all_metric_families = gather_metrics();
    let mut all_metrics = HashMap::new();
    for metric_family in all_metric_families {
        let values: Vec<_> = match metric_family.get_field_type() {
            MetricType::COUNTER => metric_family
                .get_metric()
                .iter()
                .map(|m| m.get_counter().get_value().to_string())
                .collect(),
            MetricType::GAUGE => metric_family
                .get_metric()
                .iter()
                .map(|m| m.get_gauge().get_value().to_string())
                .collect(),
            MetricType::SUMMARY => panic!("Unsupported Metric 'SUMMARY'"),
            MetricType::UNTYPED => panic!("Unsupported Metric 'UNTYPED'"),
            MetricType::HISTOGRAM => metric_family
                .get_metric()
                .iter()
                .map(|m| m.get_histogram().get_sample_count().to_string())
                .collect(),
        };
        let metric_names = metric_family.get_metric().iter().map(|m| {
            let label_strings: Vec<String> = m
                .get_label()
                .iter()
                .map(|l| format!("{}={}", l.get_name(), l.get_value()))
                .collect();
            let labels_string = format!("{{{}}}", label_strings.join(","));
            format!("{}{}", metric_family.get_name(), labels_string)
        });

        for (name, value) in metric_names.zip(values.into_iter()) {
            all_metrics.insert(name, value);
        }
    }

    all_metrics
}

// Launches a background thread which will periodically collect metrics
// every interval and write them to the provided file
pub fn dump_all_metrics_to_file_periodically<P: AsRef<Path>>(
    dir_path: &P,
    file_name: &str,
    interval: u64,
) {
    let mut file = get_metrics_file(dir_path, file_name);
    thread::spawn(move || loop {
        let mut buffer = get_all_metrics_as_serialized_string().expect("Error gathering metrics");
        if !buffer.is_empty() {
            buffer.push(b'\n');
            file.write_all(&buffer).expect("Error writing metrics");
        }
        thread::sleep(time::Duration::from_millis(interval));
    });
}

/// Helper function to record metrics for external calls.
/// Include call counts, time, and whether it's inside or not (1 or 0).
/// It assumes a OpMetrics defined as OP_COUNTERS in crate::counters;
#[macro_export]
macro_rules! monitor {
    ( $name:literal, $fn:expr ) => {{
        use crate::counters::OP_COUNTERS;
        let _timer = OP_COUNTERS.timer($name);
        let gauge = OP_COUNTERS.gauge(concat!($name, "_running"));
        gauge.inc();
        let result = $fn;
        gauge.dec();
        result
    }};
}