Provide appoximation of working set using hyper-log-log algorithm in LFC (#6935)

## Summary of changes

Calculate number of unique page accesses at compute.
It can be used to estimate working set size and adjust cache size
(shared_buffers or local file cache).

Approximation is made using HyperLogLog algorithm.
It is performed by local file cache and so is available only when local
file cache is enabled.

This calculation doesn't take in account access to the pages present in
shared buffers, but includes pages available in local file cache.

This information can be retrieved using
approximate_working_set_size(reset bool) function from neon extension.
reset parameter can be used to reset statistic and so collect unique
accesses for the particular interval.

Below is an example of estimating working set size after pgbench -c 10
-S -T 100 -s 10:
```
postgres=# select approximate_working_set_size(false);
 approximate_working_set_size 
------------------------------
                        19052
(1 row)

postgres=# select pg_table_size('pgbench_accounts')/8192;
 ?column? 
----------
    16402
(1 row)
```


## Checklist before requesting a review

- [ ] I have performed a self-review of my code.
- [ ] If it is a core feature, I have added thorough tests.
- [ ] Do we need to implement analytics? if so did you add the relevant
metrics to the dashboard?
- [ ] If this PR requires public announcement, mark it with
/release-notes label and add several sentences in this section.

## Checklist before merging

- [ ] Do not forget to reformat commit message to not include the above
checklist

---------

Co-authored-by: Konstantin Knizhnik <knizhnik@neon.tech>

pgxn/neon/Makefile

@@ -21,7 +21,7 @@ SHLIB_LINK_INTERNAL = $(libpq)
 SHLIB_LINK = -lcurl
 
 EXTENSION = neon
-DATA = neon--1.0.sql neon--1.0--1.1.sql neon--1.1--1.2.sql
+DATA = neon--1.0.sql neon--1.0--1.1.sql neon--1.1--1.2.sql neon--1.2--1.3.sql
 PGFILEDESC = "neon - cloud storage for PostgreSQL"
 
 EXTRA_CLEAN = \

pgxn/neon/file_cache.c

@@ -25,6 +25,8 @@
 #include "funcapi.h"
 #include "miscadmin.h"
 #include "pagestore_client.h"
+#include "common/hashfn.h"
+#include "lib/hyperloglog.h"
 #include "pgstat.h"
 #include "postmaster/bgworker.h"
 #include RELFILEINFO_HDR
@@ -60,6 +62,7 @@
 #define BLOCKS_PER_CHUNK	128 /* 1Mb chunk */
 #define MB					((uint64)1024*1024)
 
+#define HYPER_LOG_LOG_BIT_WIDTH   10
 #define SIZE_MB_TO_CHUNKS(size) ((uint32)((size) * MB / BLCKSZ / BLOCKS_PER_CHUNK))
 
 typedef struct FileCacheEntry
@@ -84,6 +87,8 @@ typedef struct FileCacheControl
 	uint64		writes;
 	dlist_head	lru;			/* double linked list for LRU replacement
 								 * algorithm */
+	hyperLogLogState wss_estimation; /* estimation of wroking set size */
+	uint8_t		hyperloglog_hashes[(1 << HYPER_LOG_LOG_BIT_WIDTH) + 1];
 } FileCacheControl;
 
 static HTAB *lfc_hash;
@@ -232,6 +237,14 @@ lfc_shmem_startup(void)
 		lfc_ctl->writes = 0;
 		dlist_init(&lfc_ctl->lru);
 
+		/* Initialize hyper-log-log structure for estimating working set size */
+		initHyperLogLog(&lfc_ctl->wss_estimation, HYPER_LOG_LOG_BIT_WIDTH);
+
+		/* We need hashes in shared memory */
+		pfree(lfc_ctl->wss_estimation.hashesArr);
+		memset(lfc_ctl->hyperloglog_hashes, 0, sizeof lfc_ctl->hyperloglog_hashes);
+		lfc_ctl->wss_estimation.hashesArr = lfc_ctl->hyperloglog_hashes;
+
 		/* Recreate file cache on restart */
 		fd = BasicOpenFile(lfc_path, O_RDWR | O_CREAT | O_TRUNC);
 		if (fd < 0)
@@ -529,6 +542,11 @@ lfc_read(NRelFileInfo rinfo, ForkNumber forkNum, BlockNumber blkno,
 	}
 
 	entry = hash_search_with_hash_value(lfc_hash, &tag, hash, HASH_FIND, NULL);
+
+	/* Approximate working set */
+	tag.blockNum = blkno;
+	addHyperLogLog(&lfc_ctl->wss_estimation, hash_bytes((uint8_t const*)&tag, sizeof(tag)));
+
 	if (entry == NULL || (entry->bitmap[chunk_offs >> 5] & (1 << (chunk_offs & 31))) == 0)
 	{
 		/* Page is not cached */
@@ -967,3 +985,21 @@ local_cache_pages(PG_FUNCTION_ARGS)
 	else
 		SRF_RETURN_DONE(funcctx);
 }
+
+PG_FUNCTION_INFO_V1(approximate_working_set_size);
+
+Datum
+approximate_working_set_size(PG_FUNCTION_ARGS)
+{
+	int32 dc = -1;
+	if (lfc_size_limit != 0)
+	{
+		bool reset = PG_GETARG_BOOL(0);
+		LWLockAcquire(lfc_lock, reset ? LW_EXCLUSIVE : LW_SHARED);
+		dc = (int32) estimateHyperLogLog(&lfc_ctl->wss_estimation);
+		if (reset)
+			memset(lfc_ctl->hyperloglog_hashes, 0, sizeof lfc_ctl->hyperloglog_hashes);
+		LWLockRelease(lfc_lock);
+	}
+	PG_RETURN_INT32(dc);
+}

pgxn/neon/neon--1.2--1.3.sql

@@ -0,0 +1,9 @@
+\echo Use "ALTER EXTENSION neon UPDATE TO '1.3'" to load this file. \quit
+
+CREATE FUNCTION approximate_working_set_size(reset bool)
+RETURNS integer
+AS 'MODULE_PATHNAME', 'approximate_working_set_size'
+LANGUAGE C PARALLEL SAFE;
+
+GRANT EXECUTE ON FUNCTION approximate_working_set_size(bool) TO pg_monitor;
+

pgxn/neon/neon.control

@@ -1,6 +1,6 @@
 # neon extension
 comment = 'cloud storage for PostgreSQL'
-default_version = '1.2'
+default_version = '1.3'
 module_pathname = '$libdir/neon'
 relocatable = true
 trusted = true

test_runner/regress/test_neon_extension.py

@@ -23,7 +23,7 @@ def test_neon_extension(neon_env_builder: NeonEnvBuilder):
             # IMPORTANT:
             # If the version has changed, the test should be updated.
             # Ensure that the default version is also updated in the neon.control file
-            assert cur.fetchone() == ("1.2",)
+            assert cur.fetchone() == ("1.3",)
             cur.execute("SELECT * from neon.NEON_STAT_FILE_CACHE")
             res = cur.fetchall()
             log.info(res)