rust/neon
1
0
Fork 0
mirror of https://github.com/neondatabase/neon.git synced 2026-05-19 14:10:37 +00:00
Code Issues Packages Projects Releases Wiki Activity

page cache: improve eviction-related metrics

Browse Source 
These changes help with identifying thrashing.

The existing `pageserver_page_cache_find_victim_iters_total` is already
useful, but, it doesn't tell us how many individual find_victim() calls
are happening, only how many clock-LRU steps happened in the entire system,
without info about whether we needed to actually evict other data vs
just scan for a long time, e.g., because the cache is large.

The changes in this PR allows us to
1. count each possible outcome separately, esp evictions
2. compute mean iterations/outcome

I don't think anyone except me was paying close attention to
`pageserver_page_cache_find_victim_iters_total` before, so,
I think the slight behavior change of also counting iterations
for the 'iters exceeded' case is fine.

refs https://github.com/neondatabase/cloud/issues/8351
refs https://github.com/neondatabase/neon/issues/5479
This commit is contained in:
Christian Schwarz
2023-12-08 12:17:42 +00:00
parent 79a4ee938b
commit d8c90ea659
1 changed files with 57 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

57
pageserver/src/metrics.rs
View File

@@ -337,6 +337,63 @@ pub(crate) mod page_cache_eviction_metrics {
}
}
pub(crate) mod page_cache_eviction_metrics {
use std::num::NonZeroUsize;
use metrics::{register_int_counter_vec, IntCounter, IntCounterVec};
use once_cell::sync::Lazy;
#[derive(Clone, Copy)]
pub(crate) enum Outcome {
FoundSlotUnused { iters: NonZeroUsize },
FoundSlotEvicted { iters: NonZeroUsize },
ItersExceeded { iters: NonZeroUsize },
}
static ITERS_TOTAL_VEC: Lazy<IntCounterVec> = Lazy::new(|| {
register_int_counter_vec!(
"pageserver_page_cache_find_victim_iters_total",
"Counter for the number of iterations in the find_victim loop",
&["outcome"],
)
.expect("failed to define a metric")
});
static CALLS_VEC: Lazy<IntCounterVec> = Lazy::new(|| {
register_int_counter_vec!(
"pageserver_page_cache_find_victim_calls",
"Incremented at the end of each find_victim() call.\
Filter by outcome to get e.g., eviction rate.",
&["outcome"]
)
.unwrap()
});
pub(crate) fn observe(outcome: Outcome) {
macro_rules! dry {
($label:literal, $iters:expr) => {{
static LABEL: &'static str = $label;
static ITERS_TOTAL: Lazy<IntCounter> =
Lazy::new(|| ITERS_TOTAL_VEC.with_label_values(&[LABEL]));
static CALLS: Lazy<IntCounter> =
Lazy::new(|| CALLS_VEC.with_label_values(&[LABEL]));
ITERS_TOTAL.inc_by(($iters.get()) as u64);
CALLS.inc();
}};
}
match outcome {
Outcome::FoundSlotUnused { iters } => dry!("found_empty", iters),
Outcome::FoundSlotEvicted { iters } => {
dry!("found_evicted", iters)
}
Outcome::ItersExceeded { iters } => {
dry!("err_iters_exceeded", iters);
super::page_cache_errors_inc(super::PageCacheErrorKind::EvictIterLimit);
}
}
}
}
pub(crate) static PAGE_CACHE_ACQUIRE_PINNED_SLOT_TIME: Lazy<Histogram> = Lazy::new(|| {
register_histogram!(
"pageserver_page_cache_acquire_pinned_slot_seconds",