[!] remove obsolete db_stats_aurora metric (#930)

* [!] deprecate `db_stats_aurora` metric

The `db_stats_aurora` metric was created as a workaround for early
AWS Aurora PostgreSQL limitations that prevented using certain
PostgreSQL functions like pg_backup_start_time() and complex subqueries
for invalid index detection.

Modern Aurora PostgreSQL versions (13+) now support virtually all
standard PostgreSQL administration functions, making this
Aurora-specific variant unnecessary. Users can simply use the standard
`db_stats` metric instead, which provides the same functionality
plus additional features like backup duration tracking and invalid
index monitoring.

This removal reduces maintenance overhead and eliminates confusing

* [*] remove `aurora` preset
* [*] update docs

Author: pashagolub

docs/howto/using_managed_services.md

@@ -27,58 +27,58 @@ such a use case.
 pgwatch has been tested to work with the following managed database
 services:
 
-# Google Cloud SQL for PostgreSQL
+## Google Cloud SQL for PostgreSQL
 
--   No Python / OS helpers possible. OS metrics can be integrated in
+- No Python / OS helpers possible. OS metrics can be integrated in
     Grafana though using the
     [Stackdriver](https://grafana.com/docs/grafana/latest/datasources/google-cloud-monitoring/)
     data source.
--   `pg_monitor` system role available.
--   pgwatch default preset name: `gce`.
--   Documentation: <https://cloud.google.com/sql/docs/postgres>
+- `pg_monitor` system role available.
+- pgwatch default preset name: `gce`.
+- Documentation: <https://cloud.google.com/sql/docs/postgres>
 
 To get most out pgwatch on GCE you need some additional clicks in the
 GUI / Cloud Console "Flags" section to enable some common PostgreSQL
 monitoring parameters like `track_io_timing` and `track_functions`.
 
-# Amazon RDS for PostgreSQL
+## Amazon RDS for PostgreSQL
 
--   No Python / OS helpers possible. OS metrics can be integrated in
+- No Python / OS helpers possible. OS metrics can be integrated in
     Grafana though using the
     [CloudWatch](https://grafana.com/docs/grafana/latest/datasources/cloudwatch/)
     data source
 
--   `pg_monitor` system role available.
+- `pg_monitor` system role available.
 
--   pgwatch default preset names: `rds`, `aurora`
+- pgwatch default preset names: `rds`
+
+- Documentation:
 
--   Documentation:
-    
     <https://docs.aws.amazon.com/AmazonRDS/latest/UserGuide/CHAP_PostgreSQL.html>
     <https://docs.aws.amazon.com/AmazonRDS/latest/AuroraUserGuide/Aurora.AuroraPostgreSQL.html>
 
 Note that the AWS Aurora PostgreSQL-compatible engine is missing some
 additional metrics compared to normal RDS.
 
-# Azure Database for PostgreSQL
+## Azure Database for PostgreSQL
 
--   No Python / OS helpers possible. OS metrics can be integrated in
+- No Python / OS helpers possible. OS metrics can be integrated in
     Grafana though using the [Azure
     Monitor](https://grafana.com/docs/grafana/latest/datasources/azuremonitor/)
     data source
--   `pg_monitor` system role available.
--   pgwatch default preset name: `azure`
--   Documentation: <https://docs.microsoft.com/en-us/azure/postgresql/>
+- `pg_monitor` system role available.
+- pgwatch default preset name: `azure`
+- Documentation: <https://docs.microsoft.com/en-us/azure/postgresql/>
 
 Surprisingly on Azure some file access functions are whitelisted, thus
 one can for example use the `wal_size` metrics.
 
 !!! Note
     By default Azure has **pg_stat_statements** not fully activated by
-    default, so you need to enable it manually or via the API. Documentation
-    link [here](https://docs.microsoft.com/en-us/azure/postgresql/howto-optimize-query-stats-collection).
+    default, so you need to enable it manually or via the API.
+    Documentation: <https://docs.microsoft.com/en-us/azure/postgresql/howto-optimize-query-stats-collection>
 
-# Aiven for PostgreSQL
+## Aiven for PostgreSQL
 
 The [Aiven developer
 documentation](https://aiven.io/docs/products/postgresql/howto/monitor-with-pgwatch2)

docs/reference/advanced_features.md

@@ -141,6 +141,5 @@ plain removed. Thus, to reduce server log errors and save time on
 experimenting there are following presets available:
 
 - **aws** - for standard AWS RDS managed PostgreSQL databases
-- **aurora** - for AWS Aurora managed PostgreSQL service
 - **azure** - for Azure Database for PostgreSQL managed databases
 - **gce** - for Google Cloud SQL for PostgreSQL managed databases

internal/metrics/metrics.yaml

@@ -714,45 +714,6 @@ metrics:
             - checksum_last_failure_s
             - parallel_workers_to_launch
             - parallel_workers_launched
-    db_stats_aurora:
-        description: >
-            Retrieves key statistics from the PostgreSQL `pg_stat_database` view for Amazon Aurora PostgreSQL, providing insights into the current database's performance.
-            It returns the number of backends, transaction commits and rollbacks, buffer reads and hits, tuple statistics, conflicts, temporary files and bytes,
-            deadlocks, block read and write times, postmaster uptime, recovery status, system identifier, and invalid indexes.
-            This metric helps administrators monitor database activity and performance in an Aurora PostgreSQL environment.
-        sqls:
-            11: |-
-                select /* pgwatch_generated */
-                  (extract(epoch from now()) * 1e9)::int8 as epoch_ns,
-                  numbackends,
-                  xact_commit,
-                  xact_rollback,
-                  blks_read,
-                  blks_hit,
-                  tup_returned,
-                  tup_fetched,
-                  tup_inserted,
-                  tup_updated,
-                  tup_deleted,
-                  conflicts,
-                  temp_files,
-                  temp_bytes,
-                  deadlocks,
-                  blk_read_time,
-                  blk_write_time,
-                  extract(epoch from (now() - pg_postmaster_start_time()))::int8 as postmaster_uptime_s,
-                  case when pg_is_in_recovery() then 1 else 0 end as in_recovery_int,
-                  system_identifier::text as tag_sys_id
-                from
-                  pg_stat_database, pg_control_system()
-                where
-                  datname = current_database()
-        gauges:
-            - numbackends
-            - postmaster_uptime_s
-            - backup_duration_s
-            - checksum_last_failure_s
-        storage_name: db_stats
     index_hashes:
         description: >
             Retrieves the hash of index definitions in the PostgreSQL database, providing a way to track changes in index definitions over time.
@@ -4622,29 +4583,6 @@ presets:
             table_stats: 300
             wal: 60
             wal_receiver: 120
-    aurora:
-        description: AWS Aurora doesn't expose all Postgres functions and there's no WAL
-        metrics:
-            archiver: 60
-            backends: 60
-            bgwriter: 60
-            checkpointer: 60
-            change_events: 300
-            db_size: 300
-            db_stats_aurora: 60
-            index_stats: 900
-            locks: 60
-            locks_mode: 60
-            replication: 120
-            replication_slots: 120
-            settings: 7200
-            sproc_stats: 180
-            stat_statements: 180
-            stat_statements_calls: 60
-            table_bloat_approx_summary_sql: 7200
-            table_io_stats: 600
-            table_stats: 300
-            wal_receiver: 120
     azure:
         description: similar to 'exhaustive' with stuff that's not accessible on Azure Database for PostgreSQL removed
         metrics:
@@ -4943,7 +4881,6 @@ presets:
             db_size: 30
             db_size_approx: 30
             db_stats: 30
-            db_stats_aurora: 30
             index_hashes: 30
             index_stats: 30
             instance_up: 30

internal/reaper/metric.go

@@ -13,9 +13,7 @@ import (
 const (
 	monitoredDbsDatastoreSyncIntervalSeconds = 600              // write actively monitored DBs listing to metrics store after so many seconds
 	monitoredDbsDatastoreSyncMetricName      = "configured_dbs" // FYI - for Postgres datastore there's also the admin.all_unique_dbnames table with all recent DB unique names with some metric data
-
-	dbSizeCachingInterval = 30 * time.Minute
-	dbMetricJoinStr       = "¤¤¤" // just some unlikely string for a DB name to avoid using maps of maps for DB+metric data
+	dbMetricJoinStr                          = "¤¤¤"            // just some unlikely string for a DB name to avoid using maps of maps for DB+metric data
 
 )