Files
neon/libs/utils/src/elapsed_accum.rs
Christian Schwarz aad410c8f1 improve ondemand-download latency observability (#11421)
## Problem

We don't have metrics to exactly quantify the end user impact of
on-demand downloads.

Perf tracing is underway (#11140) to supply us with high-resolution
*samples*.

But it will also be useful to have some aggregate per-timeline and
per-instance metrics that definitively contain all observations.

## Summary of changes

This PR consists of independent commits that should be reviewed
independently.

However, for convenience, we're going to merge them together.

- refactor(metrics): measure_remote_op can use async traits
- impr(pageserver metrics): task_kind dimension for
remote_timeline_client latency histo
  - implements https://github.com/neondatabase/cloud/issues/26800
- refs
https://github.com/neondatabase/cloud/issues/26193#issuecomment-2769705793
- use the opportunity to rename the metric and add a _global suffix;
checked grafana export, it's only used in two personal dashboards, one
of them mine, the other by Heikki
- log on-demand download latency for expensive-to-query but precise
ground truth
- metric for wall clock time spent waiting for on-demand downloads

## Refs

- refs https://github.com/neondatabase/cloud/issues/26800
- a bunch of minor investigations / incidents into latency outliers
2025-04-04 18:04:39 +00:00

27 lines
565 B
Rust

use std::time::{Duration, Instant};
#[derive(Default)]
pub struct ElapsedAccum {
accum: Duration,
}
impl ElapsedAccum {
pub fn get(&self) -> Duration {
self.accum
}
pub fn guard(&mut self) -> impl Drop + '_ {
let start = Instant::now();
scopeguard::guard(start, |last_wait_at| {
self.accum += Instant::now() - last_wait_at;
})
}
pub async fn measure<Fut, O>(&mut self, fut: Fut) -> O
where
Fut: Future<Output = O>,
{
let _guard = self.guard();
fut.await
}
}