[proxy] add metrics for caches (#12752)

Exposes metrics for caches. LKB-2594

This exposes a high level namespace, `cache`, that all cache metrics can
be added to - this makes it easier to make library panels for the caches
as I understand it.

To calculate the current cache fill ratio, you could use the following
query:

```
(
    cache_inserted_total{cache="node_info"}
  - sum (cache_evicted_total{cache="node_info"}) without (cause)
)
  / cache_capacity{cache="node_info"}
```

To calculate the cache hit ratio, you could use the following query:

```
  cache_request_total{cache="node_info", outcome="hit"}
/ sum (cache_request_total{cache="node_info"}) without (outcome)
```
This commit is contained in:
Conrad Ludgate
2025-07-28 11:41:49 +01:00
committed by GitHub
parent a6e0baf31a
commit effd6bf829
6 changed files with 177 additions and 29 deletions

View File

@@ -2,8 +2,12 @@ use std::ops::{Deref, DerefMut};
use std::time::{Duration, Instant};
use moka::Expiry;
use moka::notification::RemovalCause;
use crate::control_plane::messages::ControlPlaneErrorMessage;
use crate::metrics::{
CacheEviction, CacheKind, CacheOutcome, CacheOutcomeGroup, CacheRemovalCause, Metrics,
};
/// Default TTL used when caching errors from control plane.
pub const DEFAULT_ERROR_TTL: Duration = Duration::from_secs(30);
@@ -130,3 +134,35 @@ impl<K, V> Expiry<K, ControlPlaneResult<V>> for CplaneExpiry {
self.expire_early(value, updated_at)
}
}
pub fn eviction_listener(kind: CacheKind, cause: RemovalCause) {
let cause = match cause {
RemovalCause::Expired => CacheRemovalCause::Expired,
RemovalCause::Explicit => CacheRemovalCause::Explicit,
RemovalCause::Replaced => CacheRemovalCause::Replaced,
RemovalCause::Size => CacheRemovalCause::Size,
};
Metrics::get()
.cache
.evicted_total
.inc(CacheEviction { cache: kind, cause });
}
#[inline]
pub fn count_cache_outcome<T>(kind: CacheKind, cache_result: Option<T>) -> Option<T> {
let outcome = if cache_result.is_some() {
CacheOutcome::Hit
} else {
CacheOutcome::Miss
};
Metrics::get().cache.request_total.inc(CacheOutcomeGroup {
cache: kind,
outcome,
});
cache_result
}
#[inline]
pub fn count_cache_insert(kind: CacheKind) {
Metrics::get().cache.inserted_total.inc(kind);
}