Simplify the code by switching to http for IPC

Makefile.am

@@ -1,2 +1,2 @@
 bin_SCRIPTS = pgmon.py
-dist_doc_DATA = README ChangeLog pgmon_userparameter.conf pgmon_templates.yaml
+dist_doc_DATA = README ChangeLog pgmon_templates.yaml

pgmon-metrics.cfg

@@ -1,14 +0,0 @@
-# Discovery metrics
-metric=discover_dbs:set::SELECT datname, %s AS agent, %s AS cluster FROM pg_database
-metric=discover_rep:set::SELECT client_addr || '_' || regexp_replace(application_name, '[ ,]', '_', 'g') AS repid, client_addr, state, %s AS agent, %s AS cluster FROM pg_stat_replication
-
-# Cluster-wide metrics
-metric=version:value::SHOW server_version_num
-metric=max_frozen_age:value::SELECT max(age(datfrozenxid)) FROM pg_database
-
-# Per-database metrics
-metric=db_stats:row::SELECT 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 stats_reset)::float FROM pg_stat_database WHERE datname = '{datname}'
-metric=db_stats:row:140000:SELECT 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, COALESCE(checksum_failures, 0) AS checksum_failures, blk_read_time, blk_write_time, session_time, active_time, idle_in_transaction_time, sessions, sessions_abandoned, sessions_fatal, sessions_killed, extract('epoch' from stats_reset)::float FROM pg_stat_database WHERE datname = '{datname}'
-
-# Per-replication metrics
-metric=rep_stats:row::SELECT * FROM pg_stat_database WHERE client_addr || '_' || regexp_replace(application_name, '[ ,]', '_', 'g') = '{repid}'

pgmon-metrics.yml

@@ -0,0 +1,39 @@
+metrics:
+  # Discovery metrics
+  discover_dbs:
+    type: set
+    query:
+      0: SELECT datname FROM pg_database
+  discover_rep:
+    type: set
+    query:
+      0: SELECT client_addr || '_' || regexp_replace(application_name, '[ ,]', '_', 'g') AS repid, client_addr, state FROM pg_stat_replication
+
+  # cluster-wide metrics
+  version:
+    type: value
+    query:
+      0: SHOW server_version_num
+  max_frozen_age:
+    type: value
+    query:
+      0: SELECT max(age(datfrozenxid)) FROM pg_database
+
+  # Per-database metrics
+  db_stats:
+    type: row
+    query:
+      0: SELECT 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 stats_reset)::float FROM pg_stat_database WHERE datname = %(datname)s
+      140000: SELECT 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, COALESCE(checksum_failures, 0) AS checksum_failures, blk_read_time, blk_write_time, session_time, active_time, idle_in_transaction_time, sessions, sessions_abandoned, sessions_fatal, sessions_killed, extract('epoch' from stats_reset)::float FROM pg_stat_database WHERE datname = %(datname)s
+
+  # Per-replication metrics
+  rep_stats:
+    type: row
+    query:
+      0: SELECT * FROM pg_stat_database WHERE client_addr || '_' || regexp_replace(application_name, '[ ,]', '_', 'g') = '{repid}'
+
+  # Debugging
+  sleep:
+    type: value
+    query:
+      0: SELECT now(), pg_sleep(5);

pgmon.cfg

@@ -1,72 +0,0 @@
-##
-# Misc agent settings
-##
-
-# Where to write/find the agent PID
-#pid_file=/tmt/pgmon.pid
-pid_file=/run/pgmon/pgmon.pid
-
-##
-# Agent communication settings
-##
-
-# IPC socket
-#ipc_socket=/tmp/pgmon.sock
-ipc_socket=/run/pgmon/pgmon.sock
-
-# IPC communication timeout (s)
-#ipc_timeout=10
-
-# Request processing timeout (s)
-#request_timeout=10
-
-# Max size of the request queue before it blocks
-#request_queue_size=100
-
-# Max time to wait when queueing a request (s)
-#request_queue_timeout=2
-
-##
-# Agent resource settings
-##
-
-# Number of worker threads
-#worker_count=4
-
-##
-# Logging settings
-##
-
-# Log level for stderr logging (or 'off')
-stderr_log_level=debug
-
-# Log level for file logging (od 'off')
-file_log_level=off
-
-# Log file
-#log_file=pgmon.log
-
-##
-# DB connection settings
-#
-# Each cluster entry is of the form:
-#   name:address:port:dbname:user:password
-#
-# Any element other than the name can be left empty to use the defaults
-##
-
-#cluster=local:/var/run/postgresql:5432:postgres:zbx_monitor:
-cluster=pg15:localhost:54315:postgres:postgres:
-cluster=pg96:localhost:54396:postgres:postgres:
-
-# Default database to connect to when none is specified for a metric
-#dbname=postgres
-
-##
-# Monitoring configuration
-##
-
-# Metrics
-#metrics={}
-
-include=pgmon-metrics.cfg

pgmon.yml

@@ -0,0 +1,21 @@
+# IPC port
+port: 5400
+
+# Max PostgreSQL connection pool size
+#pool_size: 4
+
+# Log level for stderr logging (or 'off')
+log_level: debug
+
+# Connection string (excluding dbname)
+# This can be left empty to use the libpq defaults
+connstr: "user=postgres"
+
+# Default database to connect to when none is specified for a metric
+#dbname: postgres
+
+# Metrics
+#metrics: {}
+
+include:
+  - pgmon-metrics.yml

pgmon@.service

@@ -5,12 +5,10 @@ Description=PostgreSQL Monitoring Bridge
 After=network.target
 
 [Service]
-ExecStart=/usr/local/bin/pgmon.py --server --config /etc/pgmon/%i.cfg
-ExecReload=/user/local/bin/pgmon.py --reload --config /etc/pgmon/%i.cfg
-RuntimeDirectory=pgmon
-RuntimeDirectoryMode=0755
-PIDFile=/run/pgmon/%i.pid
+ExecStart=/usr/local/bin/pgmon.py /etc/pgmon/%i.cfg
+ExecReload=kill -HUP $MAINPID
 Restart=on-failure
+Type=exec
 
 [Install]
 WantedBy=multi-user.target

pgmon_userparameter.conf

@@ -1 +0,0 @@
-UserParameter=pgmon[*],/usr/local/bin/pgmon.py -c /etc/pgmon/$1.cfg "$1,$2,$3,$4,$5,$6,$7,$8,$9"

requirements.yml

@@ -1 +1,2 @@
-psycopg2
+psycopg[binary,pool]
+pyyaml