Combine ccp_archive_command_status queries into one query.

Add semicolons to the end of all queries.
Make ccp_replication_lag_size return the replica name for grafana dashboard legend.
DROP functions rather than CREATE OR REPLACE to avoid errors due to changes in functions.

Author: dsessler7

internal/collector/generated/pgbouncer_metrics_queries.json

@@ -4,7 +4,7 @@
 # https://github.com/open-telemetry/opentelemetry-collector-contrib/blob/-/receiver/sqlqueryreceiver#metrics-queries
 # https://github.com/CrunchyData/pgmonitor/blob/v5.1.1/sql_exporter/common/crunchy_pgbouncer_121_collector.yml
 
-  - sql: "SHOW CLIENTS"
+  - sql: "SHOW CLIENTS;"
     metrics:
       - metric_name: ccp_pgbouncer_clients_wait_seconds
         value_column: wait
@@ -15,7 +15,7 @@
   # can be NULL; the collector will warn against NULL even when not used. But it will emit
   # an error log if those columns are used.
   # The host column should always point either to pgBouncer's virtual database (the null case) or to the primary.
-  - sql: "SHOW DATABASES"
+  - sql: "SHOW DATABASES;"
     metrics:
       - metric_name: ccp_pgbouncer_databases_pool_size
         value_column: pool_size
@@ -54,14 +54,14 @@
         attribute_columns: ["name", "port", "database"]
         description: "1 if this database is currently disabled, else 0"
 
-  - sql: "SHOW LISTS"
+  - sql: "SHOW LISTS;"
     metrics:
       - metric_name: ccp_pgbouncer_lists_item_count
         value_column: items
         attribute_columns: ["list"]
         description: "Count of items registered with pgBouncer"
 
-  - sql: "SHOW POOLS"
+  - sql: "SHOW POOLS;"
     metrics:
       - metric_name: ccp_pgbouncer_pools_client_active
         value_column: cl_active
@@ -92,7 +92,7 @@
           Server connections that have been idle for more than server_check_delay,
           so they need server_check_query to run on them before they can be used again
 
-  - sql: "SHOW SERVERS"
+  - sql: "SHOW SERVERS;"
     metrics:
       - metric_name: ccp_pgbouncer_servers_close_needed
         value_column: close_needed

internal/collector/generated/postgres_5s_metrics.json

@@ -43,7 +43,14 @@
 
   - sql: >
       SELECT
-      COALESCE(EXTRACT(epoch from (CURRENT_TIMESTAMP - last_archived_time)), 0) AS seconds_since_last_archive
+        COALESCE(EXTRACT(epoch from (CURRENT_TIMESTAMP - last_archived_time)), 0) AS seconds_since_last_archive,
+        archived_count,
+        failed_count,
+        CASE
+          WHEN EXTRACT(epoch from (last_failed_time - last_archived_time)) IS NULL THEN 0
+          WHEN EXTRACT(epoch from (last_failed_time - last_archived_time)) < 0 THEN 0
+          ELSE EXTRACT(epoch from (last_failed_time - last_archived_time))
+        END AS seconds_since_last_fail
       FROM pg_catalog.pg_stat_archiver;
     metrics:
       - metric_name: ccp_archive_command_status_seconds_since_last_archive
@@ -52,35 +59,16 @@
         description: Seconds since the last successful archive operation
         static_attributes:
           server: "localhost:5432"
-
-  - sql: >
-      SELECT archived_count
-      FROM pg_catalog.pg_stat_archiver
-    metrics:
       - metric_name: ccp_archive_command_status_archived_count
         value_column: archived_count
         description: Number of WAL files that have been successfully archived
         static_attributes:
           server: "localhost:5432"
-
-  - sql: >
-      SELECT failed_count
-      FROM pg_catalog.pg_stat_archiver
-    metrics:
       - metric_name: ccp_archive_command_status_failed_count
         value_column: failed_count
         description: Number of failed attempts for archiving WAL files
         static_attributes:
           server: "localhost:5432"
-
-  - sql: >
-      SELECT CASE
-          WHEN EXTRACT(epoch from (last_failed_time - last_archived_time)) IS NULL THEN 0
-          WHEN EXTRACT(epoch from (last_failed_time - last_archived_time)) < 0 THEN 0
-          ELSE EXTRACT(epoch from (last_failed_time - last_archived_time))
-          END AS seconds_since_last_fail
-      FROM pg_catalog.pg_stat_archiver
-    metrics:
       - metric_name: ccp_archive_command_status_seconds_since_last_fail 
         value_column: seconds_since_last_fail
         description: Seconds since the last recorded failure of the archive_command
@@ -198,7 +186,7 @@
 
   - sql: >
       SELECT monitor.kdapi_scalar_bigint('cpu_request') AS request
-      , monitor.kdapi_scalar_bigint('cpu_limit') AS limit
+      , monitor.kdapi_scalar_bigint('cpu_limit') AS limit;
     metrics:
       - metric_name: ccp_nodemx_cpu_limit
         value_column: limit
@@ -297,7 +285,7 @@
                 FROM monitor.proc_mountinfo() m
                 JOIN monitor.fsinfo(m.mount_point) f USING (major_number, minor_number)
                 WHERE m.mount_point IN ('/pgdata', '/pgwal') OR
-                      m.mount_point like '/tablespaces/%'
+                      m.mount_point like '/tablespaces/%';
     metrics:
       - metric_name: ccp_nodemx_data_disk_available_bytes 
         value_column: available_bytes
@@ -469,7 +457,7 @@
         ,tx_bytes
         ,tx_packets
         ,rx_bytes
-        ,rx_packets from monitor.proc_network_stats()
+        ,rx_packets from monitor.proc_network_stats();
     metrics:
       - metric_name: ccp_nodemx_network_rx_bytes
         value_column: rx_bytes
@@ -632,6 +620,7 @@
         value_column: bytes
         value_type: double
         description: Replication lag in bytes.
+        attribute_columns: ['replica']
         static_attributes:
           server: "localhost:5432"
 

internal/collector/pgbouncer_metrics_queries.yaml

@@ -71,23 +71,25 @@ $function$;
 GRANT EXECUTE ON ALL FUNCTIONS IN SCHEMA monitor TO ccp_monitoring;
 GRANT ALL ON ALL TABLES IN SCHEMA monitor TO ccp_monitoring;
 
+DROP FUNCTION IF EXISTS get_replication_lag();
 --- get_replication_lag is used by the OTel collector.
 --- get_replication_lag is created as function, so that we can query without warning on a replica.
-CREATE OR REPLACE FUNCTION get_replication_lag() RETURNS TABLE(bytes NUMERIC) AS $$
+CREATE FUNCTION get_replication_lag() RETURNS TABLE(replica text, bytes NUMERIC) AS $$
 BEGIN
     IF pg_is_in_recovery() THEN
-        RETURN QUERY SELECT 0::NUMERIC AS bytes;
+        RETURN QUERY SELECT ''::text as replica, 0::NUMERIC AS bytes;
     ELSE
-        RETURN QUERY SELECT pg_wal_lsn_diff(sent_lsn, replay_lsn) AS bytes
+        RETURN QUERY SELECT application_name AS replica, pg_wal_lsn_diff(sent_lsn, replay_lsn) AS bytes
                      FROM pg_catalog.pg_stat_replication;
     END IF;
 END;
 $$ LANGUAGE plpgsql;
 
+DROP FUNCTION IF EXISTS get_pgbackrest_info();
 --- get_pgbackrest_info is used by the OTel collector.
 --- get_pgbackrest_info is created as a function so that no ddl runs on a replica.
 --- In the query, the --stanza argument matches DefaultStanzaName, defined in internal/pgbackrest/config.go.
-CREATE OR REPLACE FUNCTION get_pgbackrest_info()
+CREATE FUNCTION get_pgbackrest_info()
 RETURNS TABLE (
     last_diff_backup BIGINT,
     last_full_backup BIGINT,