Add telemetry for access methods

Add telemetry for tracking access methods used, number of pages for
each access method, and number of instances using each access method.

Also introduces a type-based function `ts_jsonb_set_value_by_type` that
can generate correct JSONB based on the PostgreSQL type. It will
generate "bare" values for numerics, and strings for anything else
using the output function for the type.

To test this for string values, we update `ts_jsonb_add_interval` to
use this new function, which is calling the output function for the
type, just like `ts_jsonb_set_value_by_type`.

Author: mkindahl

.github/workflows/code_style.yaml

@@ -66,7 +66,7 @@ jobs:
       - name: Run codespell
         run: |
           find . -type f \( -name "*.c" -or -name "*.h" -or -name "*.yaml" -or -name "*.sh" \) \
-            -exec codespell -L "inh,larg,inout" {} \+
+            -exec codespell -L "brin,inh,larg,inout" {} \+
 
   cc_checks:
     name: Check code formatting

.unreleased/pr_6810

@@ -0,0 +1 @@
+Implements: #6810 Add telemetry for access methods

src/jsonb_utils.c

@@ -57,34 +57,76 @@ ts_jsonb_add_str(JsonbParseState *state, const char *key, const char *value)
 	ts_jsonb_add_value(state, key, &json_value);
 }
 
+static PGFunction
+get_convert_func(Oid typeid)
+{
+	switch (typeid)
+	{
+		case INT2OID:
+			return int2_numeric;
+		case INT4OID:
+			return int4_numeric;
+		case INT8OID:
+			return int8_numeric;
+		default:
+			return NULL;
+	}
+}
+
 void
-ts_jsonb_add_int32(JsonbParseState *state, const char *key, const int32 int_value)
+ts_jsonb_set_value_by_type(JsonbValue *value, Oid typeid, Datum datum)
 {
-	Numeric value;
+	switch (typeid)
+	{
+		Oid typeOut;
+		bool isvarlena;
+		char *str;
+		PGFunction func;
+
+		case INT2OID:
+		case INT4OID:
+		case INT8OID:
+		case NUMERICOID:
+			func = get_convert_func(typeid);
+			value->type = jbvNumeric;
+			value->val.numeric = DatumGetNumeric(func ? DirectFunctionCall1(func, datum) : datum);
+			break;
+
+		default:
+			getTypeOutputInfo(typeid, &typeOut, &isvarlena);
+			str = OidOutputFunctionCall(typeOut, datum);
+			value->type = jbvString;
+			value->val.string.val = str;
+			value->val.string.len = strlen(str);
+			break;
+	}
+}
 
-	value = DatumGetNumeric(DirectFunctionCall1(int4_numeric, Int32GetDatum(int_value)));
+void
+ts_jsonb_add_int32(JsonbParseState *state, const char *key, const int32 int_value)
+{
+	JsonbValue json_value;
 
-	ts_jsonb_add_numeric(state, key, value);
+	ts_jsonb_set_value_by_type(&json_value, INT4OID, Int32GetDatum(int_value));
+	ts_jsonb_add_value(state, key, &json_value);
 }
 
 void
 ts_jsonb_add_int64(JsonbParseState *state, const char *key, const int64 int_value)
 {
-	Numeric value;
-
-	value = DatumGetNumeric(DirectFunctionCall1(int8_numeric, Int64GetDatum(int_value)));
+	JsonbValue json_value;
 
-	ts_jsonb_add_numeric(state, key, value);
+	ts_jsonb_set_value_by_type(&json_value, INT8OID, Int64GetDatum(int_value));
+	ts_jsonb_add_value(state, key, &json_value);
 }
 
 void
 ts_jsonb_add_interval(JsonbParseState *state, const char *key, Interval *interval)
 {
-	char *value;
-
-	value = DatumGetCString(DirectFunctionCall1(interval_out, IntervalPGetDatum(interval)));
+	JsonbValue json_value;
 
-	ts_jsonb_add_str(state, key, value);
+	ts_jsonb_set_value_by_type(&json_value, INTERVALOID, IntervalPGetDatum(interval));
+	ts_jsonb_add_value(state, key, &json_value);
 }
 
 void

src/jsonb_utils.h

@@ -23,6 +23,7 @@ extern TSDLLEXPORT void ts_jsonb_add_int64(JsonbParseState *state, const char *k
 										   const int64 value);
 extern TSDLLEXPORT void ts_jsonb_add_numeric(JsonbParseState *state, const char *key,
 											 const Numeric value);
+extern TSDLLEXPORT void ts_jsonb_set_value_by_type(JsonbValue *value, Oid typeid, Datum datum);
 
 extern void ts_jsonb_add_value(JsonbParseState *state, const char *key, JsonbValue *value);
 

src/telemetry/telemetry.c

@@ -757,6 +757,87 @@ add_replication_telemetry(JsonbParseState *state)
 #define REQ_RELS_CONTINUOUS_AGGS "continuous_aggregates"
 #define REQ_FUNCTIONS_USED "functions_used"
 #define REQ_REPLICATION "replication"
+#define REQ_ACCESS_METHODS "access_methods"
+
+/*
+ * Add the result of a query as a sub-object to the JSONB.
+ *
+ * Each row from the query generates a separate object keyed by one of the
+ * columns. Each row will be represented as an object and stored under the
+ * "key" column. For example, with this query:
+ *
+ *    select amname as name,
+ *           sum(relpages) as pages,
+ *			 count(*) as instances
+ *		from pg_class join pg_am on relam = pg_am.oid
+ *	  group by pg_am.oid;
+ *
+ * might generate the object
+ *
+ * {
+ *    "brin" : {
+ *       "instances" : 44,
+ *       "pages" : 432
+ *    },
+ *    "btree" : {
+ *       "instances" : 99,
+ *       "pages" : 1234
+ *    }
+ * }
+ */
+static void
+add_query_result_dict(JsonbParseState *state, const char *query)
+{
+	MemoryContext orig_context = CurrentMemoryContext;
+
+	int res;
+	if (SPI_connect() != SPI_OK_CONNECT)
+		elog(ERROR, "could not connect to SPI");
+
+	/* Lock down search_path */
+	res = SPI_execute("SET LOCAL search_path TO pg_catalog, pg_temp", false, 0);
+	Ensure(res >= 0, "could not set search path");
+
+	res = SPI_execute(query, true, 0);
+	Ensure(res >= 0, "could not execute query");
+
+	MemoryContext spi_context = MemoryContextSwitchTo(orig_context);
+
+	(void) pushJsonbValue(&state, WJB_BEGIN_OBJECT, NULL);
+	for (uint64 r = 0; r < SPI_processed; r++)
+	{
+		char *key_string = SPI_getvalue(SPI_tuptable->vals[r], SPI_tuptable->tupdesc, 1);
+		JsonbValue key = {
+			.type = jbvString,
+			.val.string.val = pstrdup(key_string),
+			.val.string.len = strlen(key_string),
+		};
+
+		(void) pushJsonbValue(&state, WJB_KEY, &key);
+
+		(void) pushJsonbValue(&state, WJB_BEGIN_OBJECT, NULL);
+		for (int c = 1; c < SPI_tuptable->tupdesc->natts; ++c)
+		{
+			bool isnull;
+			Datum val_datum =
+				SPI_getbinval(SPI_tuptable->vals[r], SPI_tuptable->tupdesc, c + 1, &isnull);
+			if (!isnull)
+			{
+				char *key_string = SPI_fname(SPI_tuptable->tupdesc, c + 1);
+				JsonbValue value;
+				ts_jsonb_set_value_by_type(&value,
+										   SPI_gettypeid(SPI_tuptable->tupdesc, c + 1),
+										   val_datum);
+				ts_jsonb_add_value(state, key_string, &value);
+			}
+		}
+		pushJsonbValue(&state, WJB_END_OBJECT, NULL);
+	}
+	MemoryContextSwitchTo(spi_context);
+	res = SPI_finish();
+	Assert(res == SPI_OK_FINISH);
+	(void) pushJsonbValue(&state, WJB_END_OBJECT, NULL);
+}
 
 static Jsonb *
 build_telemetry_report()
@@ -937,6 +1018,15 @@ build_telemetry_report()
 	add_replication_telemetry(parse_state);
 	pushJsonbValue(&parse_state, WJB_END_OBJECT, NULL);
 
+	key.type = jbvString;
+	key.val.string.val = REQ_ACCESS_METHODS;
+	key.val.string.len = strlen(REQ_ACCESS_METHODS);
+	(void) pushJsonbValue(&parse_state, WJB_KEY, &key);
+	add_query_result_dict(parse_state,
+						  "SELECT amname AS name, sum(relpages) AS pages, count(*) AS "
+						  "instances FROM pg_class JOIN pg_am ON relam = pg_am.oid "
+						  "GROUP BY amname");
+
 	/* end of telemetry object */
 	result = pushJsonbValue(&parse_state, WJB_END_OBJECT, NULL);
 

test/expected/telemetry.out

@@ -350,6 +350,7 @@ WHERE key != 'os_name_pretty';
  db_metadata
  replication
  build_os_name
+ access_methods
  functions_used
  install_method
  installed_time
@@ -377,7 +378,7 @@ WHERE key != 'os_name_pretty';
  num_compression_policies_fixed
  num_user_defined_actions_fixed
  num_continuous_aggs_policies_fixed
-(37 rows)
+(38 rows)
 
 CREATE MATERIALIZED VIEW telemetry_report AS
 SELECT t FROM get_telemetry_report() t;
@@ -395,6 +396,17 @@ SELECT t -> 'instance_metadata' FROM telemetry_report;
  {"cloud": "ci"}
 (1 row)
 
+-- Check access methods
+SELECT t->'access_methods' ? 'btree',
+       t->'access_methods' ? 'heap',
+       CAST(t->'access_methods'->'btree'->'pages' AS int) > 0,
+       CAST(t->'access_methods'->'btree'->'instances' AS int) > 0
+  FROM telemetry_report;
+ ?column? | ?column? | ?column? | ?column? 
+----------+----------+----------+----------
+ t        | t        | t        | t
+(1 row)
+
 WITH t AS (
 	 SELECT t -> 'relations' AS rels
 	 FROM telemetry_report

test/sql/telemetry.sql

@@ -166,6 +166,13 @@ SELECT t -> 'db_metadata' FROM telemetry_report;
 -- check timescaledb_telemetry.cloud
 SELECT t -> 'instance_metadata' FROM telemetry_report;
 
+-- Check access methods
+SELECT t->'access_methods' ? 'btree',
+       t->'access_methods' ? 'heap',
+       CAST(t->'access_methods'->'btree'->'pages' AS int) > 0,
+       CAST(t->'access_methods'->'btree'->'instances' AS int) > 0
+  FROM telemetry_report;
+
 WITH t AS (
 	 SELECT t -> 'relations' AS rels
 	 FROM telemetry_report