Add metrics for cache entry memory size

.changesets/feat_enhanced_observability.md

@@ -0,0 +1,13 @@
+### New `apollo.router.cache.storage.estimated_size` gauge ([PR #5770](https://github.com/apollographql/router/pull/5770))
+
+The router supports the new metric `apollo.router.cache.storage.estimated_size` that helps users understand and monitor the amount of memory that query planner cache entries consume.
+
+The `apollo.router.cache.storage.estimated_size` metric gives an estimated size in bytes of a cache entry. It has the following attributes:
+- `kind`: `query planner`.
+- `storage`: `memory`.
+
+Before using the estimate to decide whether to update the cache, users should validate that the estimate correlates with their pod's memory usage. 
+
+To learn how to troubleshoot with this metric, see the [Pods terminating due to memory pressure](https://www.apollographql.com/docs/router/containerization/kubernetes#pods-terminating-due-to-memory-pressure) guide in docs.
+
+By [@BrynCooke](https://github.com/BrynCooke) in https://github.com/apollographql/router/pull/5770

.changesets/fix_missing_cache_gauge.md

@@ -0,0 +1,5 @@
+### Fix missing `apollo_router_cache_size` metric ([PR #5770](https://github.com/apollographql/router/pull/5770))
+
+Previously, if the in-memory cache wasn't mutated, the `apollo_router_cache_size` metric wouldn't be available. This has been fixed in this release.
+
+By [@BrynCooke](https://github.com/BrynCooke) in https://github.com/apollographql/router/pull/5770

apollo-router/src/cache/mod.rs

@@ -14,7 +14,9 @@ use self::storage::ValueType;
 use crate::configuration::RedisCache;
 
 pub(crate) mod redis;
+mod size_estimation;
 pub(crate) mod storage;
+pub(crate) use size_estimation::estimate_size;
 
 type WaitMap<K, V> = Arc<Mutex<HashMap<K, broadcast::Sender<V>>>>;
 pub(crate) const DEFAULT_CACHE_CAPACITY: NonZeroUsize = match NonZeroUsize::new(512) {
@@ -37,7 +39,7 @@ where
     pub(crate) async fn with_capacity(
         capacity: NonZeroUsize,
         redis: Option<RedisCache>,
-        caller: &str,
+        caller: &'static str,
     ) -> Result<Self, BoxError> {
         Ok(Self {
             wait_map: Arc::new(Mutex::new(HashMap::new())),
@@ -47,7 +49,7 @@ where
 
     pub(crate) async fn from_configuration(
         config: &crate::configuration::Cache,
-        caller: &str,
+        caller: &'static str,
     ) -> Result<Self, BoxError> {
         Self::with_capacity(config.in_memory.limit, config.redis.clone(), caller).await
     }

apollo-router/src/cache/redis.rs

@@ -593,12 +593,19 @@ mod test {
 
     use url::Url;
 
+    use crate::cache::storage::ValueType;
+
     #[test]
     fn ensure_invalid_payload_serialization_doesnt_fail() {
         #[derive(Clone, Debug, serde::Serialize, serde::Deserialize)]
         struct Stuff {
             time: SystemTime,
         }
+        impl ValueType for Stuff {
+            fn estimated_size(&self) -> Option<usize> {
+                None
+            }
+        }
 
         let invalid_json_payload = super::RedisValue(Stuff {
             // this systemtime is invalid, serialization will fail