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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379
// Copyright 2020 Twitter, Inc.
// Licensed under the Apache License, Version 2.0
// http://www.apache.org/licenses/LICENSE-2.0
use crate::Error;
use crate::*;
use core::sync::atomic::*;
use histogram::{Bucket, Histogram};
use parking_lot::Mutex;
/// A `Heatmap` stores counts for timestamped values over a configured span of
/// time.
///
/// Internally, it is represented as a ring buffer of histograms with one
/// additional histogram to track all counts within the span of time. Old
/// histograms age-out as time moves forward and they are subtracted from the
/// summary histogram at that point.
///
/// This acts as a moving histogram, such that requesting a percentile returns
/// a percentile from across the configured span of time.
pub struct Heatmap {
slices: Vec<Histogram>,
current: AtomicUsize,
lock: Mutex<()>,
next_tick: AtomicInstant,
resolution: Duration,
summary: Histogram,
}
/// A `Builder` allows for constructing a `Heatmap` with the desired
/// configuration.
pub struct Builder {
// minimum resolution parameter `M = 2^m`
m: u32,
// minimum resolution range parameter `R = 2^r - 1`
r: u32,
// maximum value parameter `N = 2^n - 1`
n: u32,
// span of time represented by the heatmap
span: Duration,
// the resolution in the time domain
resolution: Duration,
}
impl Builder {
/// Consume the `Builder` and return a `Heatmap`.
pub fn build(self) -> Result<Heatmap, Error> {
Heatmap::new(self.m, self.r, self.n, self.span, self.resolution)
}
/// Sets the width of the smallest bucket in the `Heatmap`.
///
/// As the `Heatmap` uses base-2 internally, the resolution will be the
/// largest power of two that is less than or equal to the provided value.
/// For example, if the minimum resolution is set to 10, the width of the
/// smallest bucket will be 8.
pub fn min_resolution(mut self, width: u64) -> Self {
self.m = 64 - width.leading_zeros();
self
}
/// Sets the maximum value that the minimum resolution extends to.
///
/// This value should be greater than the minimum resolution. If the value
/// provided is not a power of two, the smallest power of two that is larger
/// than the provided value will be used.
pub fn min_resolution_range(mut self, value: u64) -> Self {
self.r = 64 - value.next_power_of_two().leading_zeros();
self
}
/// Sets the maximum value that can be recorded into the `Heatmap`.
///
/// If the value provided is not a power of two, the smallest power of two
/// that is larger than the provided value will be used.
pub fn maximum_value(mut self, value: u64) -> Self {
self.n = 64 - value.next_power_of_two().leading_zeros();
self
}
/// Sets the duration that is covered by the `Heatmap`.
///
/// Values that are older than the duration will be dropped as they age-out.
pub fn span(mut self, duration: Duration) -> Self {
self.span = duration;
self
}
/// Sets the resolution in the time domain.
///
/// Increments with similar timestamps will be grouped together and age-out
/// together.
pub fn resolution(mut self, duration: Duration) -> Self {
self.resolution = duration;
self
}
}
impl Heatmap {
/// Create a new `Heatmap` which stores counts for timestamped values over
/// a configured span of time.
///
/// - `m` - sets the minimum resolution `M = 2^m`. This is the smallest unit
/// of quantification, which is also the smallest bucket width. If the input
/// values are always integers, choosing `m=0` would ensure precise
/// recording for the smallest values.
///
/// - `r` - sets the minimum resolution range `R = 2^r - 1`. The selected
/// value must be greater than the minimum resolution `m`. This sets the
/// maximum value that the minimum resolution should extend to.
///
/// - `n` - sets the maximum value `N = 2^n - 1`. The selected value must be
/// greater than or equal to the minimum resolution range `r`.
///
/// - `span` - sets the total duration that the heatmap covers
///
/// - `resolution` - sets the resolution in the time domain. Counts from
/// similar instants in time will be grouped together.
pub fn new(
m: u32,
r: u32,
n: u32,
span: Duration,
resolution: Duration,
) -> Result<Self, Error> {
let mut slices = Vec::new();
let mut true_span = Duration::from_nanos(0);
while true_span < span {
slices.push(Histogram::new(m, r, n)?);
true_span += resolution;
}
// allocate one extra histogram so we always have a cleared
// one in the ring
slices.push(Histogram::new(m, r, n)?);
slices.shrink_to_fit();
let next_tick = AtomicInstant::now();
next_tick.fetch_add(resolution, Ordering::Relaxed);
Ok(Self {
slices,
current: AtomicUsize::new(0),
lock: Mutex::new(()),
next_tick,
resolution,
summary: Histogram::new(m, r, n)?,
})
}
/// Creates a `Builder` with the default values `m = 0`, `r = 10`, `n = 30`,
/// `span = 60s`, `resolution = 1s`.
///
/// This would create a `Heatmap` with 61 total `Histogram`s, each with
/// 11264 buckets which can store values from 1 to 1_073_741_823 with
/// values 1 to 1023 being stored in buckets with a width of 1. Such a
/// `Heatmap` would be appropriate for latencies measured in nanoseconds
/// where the max expected latency is one second and reporting covers the
/// past minute.
pub fn builder() -> Builder {
Builder {
m: 0,
r: 10,
n: 30,
span: Duration::from_secs(60),
resolution: Duration::from_secs(1),
}
}
/// Returns the number of windows stored in the `Heatmap`
pub fn windows(&self) -> usize {
self.slices.len()
}
/// Returns the number of buckets stored within each `Histogram` in the
/// `Heatmap`
pub fn buckets(&self) -> usize {
self.summary.buckets()
}
/// Increment a time-value pair by a specified count
pub fn increment(&self, time: Instant, value: u64, count: u32) {
self.tick(time);
let _ = self.summary.increment(value, count);
let _ = self.slices[self.current.load(Ordering::Relaxed)].increment(value, count);
}
/// Return the nearest value for the requested percentile (0.0 - 100.0)
/// across the total range of samples retained in the `Heatmap`.
///
/// Note: since the heatmap stores a distribution across a configured time
/// span, sequential calls to fetch the percentile might result in different
/// results even without concurrent writers. For instance, you may see a
/// 90th percentile that is higher than the 100th percentile depending on
/// the timing of calls to this function and the distribution of your data.
///
/// Note: concurrent writes may also effect the value returned by this
/// function. Users needing better consistency should ensure that other
/// threads are not writing into the heatmap while this function is
/// in-progress.
pub fn percentile(&self, percentile: f64) -> Result<Bucket, Error> {
self.tick(Instant::now());
self.summary.percentile(percentile).map_err(Error::from)
}
/// Creates an iterator to iterate over the component histograms of this
/// heatmap.
pub fn iter(&self) -> Iter {
self.into_iter()
}
/// Access the summary histogram of this heatmap.
///
/// Note that concurrent modifications to the heatmap will continue to show
/// up in the summary histogram while it is being read so sequential
/// queries may not return consistent results.
pub fn summary(&self) -> &Histogram {
&self.summary
}
// Internal function which handles reuse of older windows to store newer
/// values.
fn tick(&self, time: Instant) {
loop {
// quick check to see if this data point requires ticking forward
let next_tick = self.next_tick.load(Ordering::Relaxed);
if time < next_tick {
return;
} else {
// some expiration needs to happen, let's try to acquire the lock
//
// note: we use parking_lot mutex as it will not be poisoned by
// a thread panic while locked.
if let Some(_lock) = self.lock.try_lock() {
// now that we have the lock, check that we still need to
// tick forward
if time < self.next_tick.load(Ordering::Relaxed) {
return;
}
// calculate the index of the next current slice
let current = self.current.load(Ordering::Relaxed);
let mut next = current + 1;
if next >= self.slices.len() {
next -= self.slices.len();
}
// move current and next_tick forward
self.current.store(next, Ordering::Relaxed);
self.next_tick.fetch_add(self.resolution, Ordering::Relaxed);
// now we have a slice to subtract and clear from the summary
// this is the histogram that is one ahead of our new current
// position
let mut to_clear = next + 1;
// check if we need to wrap around to the start
if to_clear >= self.slices.len() {
to_clear -= self.slices.len();
}
// subtract and clear
let _ = self.summary.subtract_and_clear(&self.slices[to_clear]);
}
// if we failed to acquire the lock, just loop. this does mean
// we busy wait if the heatmap has fallen behind by multiple
// ticks. we expect the typical case to be that we need to tick
// forward by just a single slice. in that case, if we fail to
// acquire the lock, we expect that the loop will terminate when
// we check `next_tick` at the start of the next iteration.
}
}
}
/// Internal function to return a `Window` from the `Heatmap`.
fn get_slice(&self, index: usize) -> Option<Window> {
if let Some(histogram) = self.slices.get(index) {
let shift = if index > self.current.load(Ordering::Relaxed) {
self.resolution.mul_f64(
(self.slices.len() + self.current.load(Ordering::Relaxed) - index) as f64,
)
} else {
self.resolution
.mul_f64((self.current.load(Ordering::Relaxed) - index) as f64)
};
Some(Window {
start: self.next_tick.load(Ordering::Relaxed) - shift - self.resolution,
stop: self.next_tick.load(Ordering::Relaxed) - shift,
histogram,
})
} else {
None
}
}
}
impl Clone for Heatmap {
fn clone(&self) -> Self {
let slices = self.slices.clone();
let summary = self.summary.clone();
let resolution = self.resolution;
let current = AtomicUsize::new(self.current.load(Ordering::Relaxed));
let next_tick = AtomicInstant::new(self.next_tick.load(Ordering::Relaxed));
Heatmap {
slices,
current,
lock: Mutex::new(()),
next_tick,
resolution,
summary,
}
}
}
pub struct Iter<'a> {
inner: &'a Heatmap,
index: usize,
visited: usize,
}
impl<'a> Iter<'a> {
fn new(inner: &'a Heatmap) -> Iter<'a> {
let index = if inner.current.load(Ordering::Relaxed) < (inner.slices.len() - 1) {
inner.current.load(Ordering::Relaxed) + 1
} else {
0
};
Iter {
inner,
index,
visited: 0,
}
}
}
impl<'a> Iterator for Iter<'a> {
type Item = Window<'a>;
fn next(&mut self) -> Option<Window<'a>> {
if self.visited >= self.inner.slices.len() {
None
} else {
let bucket = self.inner.get_slice(self.index);
self.index += 1;
if self.index >= self.inner.slices.len() {
self.index = 0;
}
self.visited += 1;
bucket
}
}
}
impl<'a> IntoIterator for &'a Heatmap {
type Item = Window<'a>;
type IntoIter = Iter<'a>;
fn into_iter(self) -> Self::IntoIter {
Iter::new(self)
}
}
#[cfg(test)]
mod tests {
use super::*;
#[test]
fn age_out() {
let heatmap =
Heatmap::new(0, 4, 20, Duration::from_secs(1), Duration::from_millis(1)).unwrap();
assert_eq!(heatmap.percentile(0.0).map(|v| v.high()), Err(Error::Empty));
heatmap.increment(Instant::now(), 1, 1);
assert_eq!(heatmap.percentile(0.0).map(|v| v.high()), Ok(1));
std::thread::sleep(std::time::Duration::from_millis(100));
assert_eq!(heatmap.percentile(0.0).map(|v| v.high()), Ok(1));
std::thread::sleep(std::time::Duration::from_millis(2000));
assert_eq!(heatmap.percentile(0.0).map(|v| v.high()), Err(Error::Empty));
}
}