Reimplement prepared statements with LRU cache and statement deduplication (#618)

* Initial commit

* Cleanup and add stats

* Use an arc instead of full clones to store the parse packets

* Use mutex instead

* fmt

* clippy

* fmt

* fix?

* fix?

* fmt

* typo

* Update docs

* Refactor custom protocol

* fmt

* move custom protocol handling to before parsing

* Support describe

* Add LRU for server side statement cache

* rename variable

* Refactoring

* Move docs

* Fix test

* fix

* Update tests

* trigger build

* Add more tests

* Reorder handling sync

* Support when a named describe is sent along with Parse (go pgx) and expecting results

* don't talk to client if not needed when client sends Parse

* fmt :(

* refactor tests

* nit

* Reduce hashing

* Reducing work done to decode describe and parse messages

* minor refactor

* Merge branch 'main' into zain/reimplment-prepared-statements-with-global-lru-cache

* Rewrite extended and prepared protocol message handling to better support mocking response packets and close

* An attempt to better handle if there are DDL changes that might break cached plans with ideas about how to further improve it

* fix

* Minor stats fixed and cleanup

* Cosmetic fixes (#64)

* Cosmetic fixes

* fix test

* Change server drop for statement cache error to a `deallocate all`

* Updated comments and added new idea for handling DDL changes impacting cached plans

* fix test?

* Revert test change

* trigger build, flakey test

* Avoid potential race conditions by changing get_or_insert to promote for pool LRU

* remove ps enabled variable on the server in favor of using an option

* Add close to the Extended Protocol buffer

---------

Co-authored-by: Lev Kokotov <levkk@users.noreply.github.com>

Author: zainkabani

CONFIG.md

@@ -259,22 +259,6 @@ Password to be used for connecting to servers to obtain the hash used for md5 au
 specified in `auth_query_user`. The connection will be established using the database configured in the pool.
 This parameter is inherited by every pool and can be redefined in pool configuration.
 
-### prepared_statements
-```
-path: general.prepared_statements
-default: false
-```
-
-Whether to use prepared statements or not.
-
-### prepared_statements_cache_size
-```
-path: general.prepared_statements_cache_size
-default: 500
-```
-
-Size of the prepared statements cache.
-
 ### dns_cache_enabled
 ```
 path: general.dns_cache_enabled
@@ -324,6 +308,15 @@ If the client doesn't specify, PgCat routes traffic to this role by default.
 `replica` round-robin between replicas only without touching the primary,
 `primary` all queries go to the primary unless otherwise specified.
 
+### prepared_statements_cache_size
+```
+path: general.prepared_statements_cache_size
+default: 0
+```
+
+Size of the prepared statements cache. 0 means disabled.
+TODO: update documentation
+
 ### query_parser_enabled
 ```
 path: pools.<pool_name>.query_parser_enabled

Cargo.lock

@@ -48,6 +48,7 @@ itertools = "0.10"
 clap = { version = "4.3.1", features = ["derive", "env"] }
 tracing = "0.1.37"
 tracing-subscriber = { version = "0.3.17", features = ["json", "env-filter", "std"]}
+lru = "0.12.0"
 
 [target.'cfg(not(target_env = "msvc"))'.dependencies]
 jemallocator = "0.5.0"

Cargo.toml

@@ -60,12 +60,6 @@ tcp_keepalives_count = 5
 # Number of seconds between keepalive packets.
 tcp_keepalives_interval = 5
 
-# Handle prepared statements.
-prepared_statements = true
-
-# Prepared statements server cache size.
-prepared_statements_cache_size = 500
-
 # Path to TLS Certificate file to use for TLS connections
 # tls_certificate = ".circleci/server.cert"
 # Path to TLS private key file to use for TLS connections
@@ -156,6 +150,10 @@ load_balancing_mode = "random"
 # `primary` all queries go to the primary unless otherwise specified.
 default_role = "any"
 
+# Prepared statements cache size.
+# TODO: update documentation
+prepared_statements_cache_size = 500
+
 # If Query Parser is enabled, we'll attempt to parse
 # every incoming query to determine if it's a read or a write.
 # If it's a read query, we'll direct it to a replica. Otherwise, if it's a write,

pgcat.toml

@@ -744,6 +744,7 @@ where
         ("age_seconds", DataType::Numeric),
         ("prepare_cache_hit", DataType::Numeric),
         ("prepare_cache_miss", DataType::Numeric),
+        ("prepare_cache_eviction", DataType::Numeric),
         ("prepare_cache_size", DataType::Numeric),
     ];
 
@@ -776,6 +777,10 @@ where
                 .prepared_miss_count
                 .load(Ordering::Relaxed)
                 .to_string(),
+            server
+                .prepared_eviction_count
+                .load(Ordering::Relaxed)
+                .to_string(),
             server
                 .prepared_cache_size
                 .load(Ordering::Relaxed)