PostgreSQL Setup

Logstag monitors PostgreSQL through system statistics views, catalog metadata, role metadata, replication metadata, and query statistics when pg_stat_statements is enabled.

Use a dedicated monitoring user instead of a personal administrator account or application user.

Supported Target Configuration

[targets.production-postgres]
platform = "self-hosted"
db_engine = "postgresql"
db_host = "postgres.example.internal"
db_port = 5432
db_username = "logstag_monitor"
db_password = "<your-database-password>"
db_name = "postgres"

For AWS RDS or Huawei Cloud RDS, set platform to the relevant managed platform value and add the cloud fields described in Managed Targets.

Access Model

For PostgreSQL 10 and later, pg_monitor is the recommended starting point. It gives the monitoring user read access to PostgreSQL statistics and settings without requiring superuser privileges.

Logstag also needs schema visibility for the schemas you want represented in schema, table, index, function, sequence, and permission views.

Access	Purpose
CONNECT on the monitored database	Allows the agent to connect.
pg_monitor	Allows access to PostgreSQL monitoring views and settings.
USAGE on monitored schemas	Allows schema metadata collection for user-defined objects.
pg_stat_statements	Enables query performance statistics when installed and loaded.

Create the Monitoring User

Run as a PostgreSQL administrator:

CREATE USER logstag_monitor WITH PASSWORD '<strong-password>';

Grant connection access to each monitored database:

GRANT CONNECT ON DATABASE <database_name> TO logstag_monitor;

Grant the monitoring role:

GRANT pg_monitor TO logstag_monitor;

Grant schema visibility for each application schema that should appear in Logstag:

GRANT USAGE ON SCHEMA <schema_name> TO logstag_monitor;

Repeat the schema grant for each schema you want monitored.

Enable Query Statistics

Logstag uses pg_stat_statements for query performance metrics. Without it, PostgreSQL monitoring still works, but query statistics are unavailable.

Enable the extension in PostgreSQL configuration:

shared_preload_libraries = 'pg_stat_statements'

This setting requires a PostgreSQL restart when it is changed.

Create the extension in each monitored database where query statistics should be collected:

CREATE EXTENSION IF NOT EXISTS pg_stat_statements;

The extension records normalized query text and execution statistics. Treat query statistics as sensitive operational metadata.

Optional PostgreSQL Settings

Some PostgreSQL settings improve metric quality:

Setting	Effect
track_io_timing	Enables block read/write timing metrics where PostgreSQL exposes them.
track_activity_query_size	Controls how much active query text PostgreSQL keeps visible in activity views.

Review these settings with the database owner before changing them. They affect PostgreSQL behavior outside Logstag.

Data Collected

Depending on server role, PostgreSQL version, enabled extensions, and permissions, Logstag can collect:

• Active sessions, wait events, blocking relationships, and client metadata.
• Database activity counters, deadlocks, temporary file usage, buffer activity, and transaction counters.
• Query performance statistics from pg_stat_statements.
• Replication lag on primary servers.
• Recovery conflict statistics on replica servers.
• Schema, table, column, index, view, function, sequence, foreign key, and constraint metadata.
• Role, membership, schema privilege, table privilege, and function privilege metadata.
• PostgreSQL settings and vacuum activity.

Query text, role names, schema names, object names, client addresses, and application names can be sensitive. Review Data Collection and Privacy before enabling production monitoring.

Primary and Replica Behavior

Logstag adjusts PostgreSQL collection based on server role:

Server Role	Collection Notes
Primary	Collects runtime activity, query statistics, database activity, schema metadata, settings, roles, and replication lag for connected replicas.
Replica	Collects runtime activity, database activity, schema metadata, settings, roles, and recovery conflict statistics.

Query statistics depend on pg_stat_statements availability in the monitored database.

Managed PostgreSQL

For managed PostgreSQL services such as AWS RDS or Huawei Cloud RDS:

• Use the managed platform value only when cloud-side system metrics should be collected.
• Confirm the database user has the PostgreSQL permissions described above.
• Confirm cloud credentials and instance metadata are configured for managed platform metrics.
• Confirm provider-side monitoring prerequisites are enabled where required.

Database-level PostgreSQL permissions and cloud-provider permissions are separate. Both must be correct for full coverage.

Validate Setup

After updating the database and agent configuration:

1. Validate the agent configuration.
2. Restart the agent.
3. Check the agent logs for connection, authentication, permission, or extension errors.
4. Confirm the PostgreSQL target appears in Logstag.
5. Confirm runtime metrics appear.
6. Confirm query statistics appear if pg_stat_statements was enabled.
7. Confirm schema and index metadata appear after the schema collection interval.

Linux:

/opt/logstag-agent/bin/logstag-agent --check-config
sudo systemctl restart logstag-agent
sudo journalctl -u logstag-agent -f

Windows:

& "C:\Program Files\Logstag Agent\bin\logstag-agent.exe" --check-config
Restart-Service "Logstag Agent"
Get-Content "C:\ProgramData\Logstag Agent\logs\agent.log" -Tail 50

Troubleshooting

If PostgreSQL data does not appear:

• Confirm the agent host can reach db_host and db_port.
• Confirm db_engine is set to postgresql.
• Confirm db_name matches the monitored database.
• Confirm the monitoring user can connect to the database.
• Confirm the monitoring user has pg_monitor.
• Confirm the monitoring user has USAGE on the schemas you expect to monitor.
• Confirm pg_stat_statements is installed and loaded if query statistics are expected.
• Confirm the agent was restarted after configuration changes.

If only query statistics are missing, focus on pg_stat_statements installation, shared_preload_libraries, database-level extension creation, and PostgreSQL restart status.