Skip to main content

Verify Agent Status

After deploying the agent, use the following methods to confirm it is running correctly and communicating with the platform.

Method 1 — Check the systemd Service​

systemctl status pmp4pg-agent

A healthy agent shows:

● pmp4pg-agent.service - PMP4PG Monitoring Agent
     Loaded: loaded (/etc/systemd/system/pmp4pg-agent.service; enabled)
     Active: active (running) since 2025-01-15 10:23:11 UTC; 5min ago
   Main PID: 12345 (pmp4pg-agent)

If the service shows failed or inactive, check the logs:

journalctl -u pmp4pg-agent -n 50 --no-pager

Method 2 — Check the Agent Logs​

# Live log stream
journalctl -u pmp4pg-agent -f

# Or from the log file directly
tail -f /var/log/pmp4pg-agent/agent.log

Signs of a healthy agent in the logs:

INFO  Heartbeat sent successfully
INFO  ASH batch sent: 50 samples
INFO  pg_stat_statements delta sent: 12 statements
INFO  pg_stat_database delta sent
INFO  Database sizes sent: 3 databases

Signs of a problem:

ERROR Failed to connect to PostgreSQL: ...
ERROR Failed to send ASH batch: connection refused
WARN  Retrying in 10s (attempt 2/3)

Method 3 — Check the Platform Dashboard​

Navigate to the PMP4PG Global Dashboard in your browser:

https://pmp.data-resilience.fr

The monitored server should appear with:

  • Status ACTIVE (green) — agent is running and sending heartbeats
  • Last heartbeat timestamp updated within the last minute
  • Agent version displayed in the server detail page

{screenshot: server-active-status-dashboard}

info

It may take up to 30 seconds after agent startup for the status to switch from INACTIVE to ACTIVE in the dashboard (heartbeat interval).

Method 4 — Server Detail Page​

Click on the server in the dashboard to open the Server Detail page. The Agent Status tab shows:

FieldExpected Value
Agent StatusACTIVE
Last HeartbeatLess than 1 minute ago
Agent Version1.0.0
Agent IDUUID matching your config.yml
CollectorsAll enabled collectors listed as running

{screenshot: server-detail-agent-status-tab}

Common Issues​

Agent starts but status stays INACTIVE in the dashboard​

  • Verify the central.api_url in config.yml is correct and reachable from the agent host
  • Check for firewall rules blocking outbound connections from the agent to the backend port
  • Confirm agent.id and central.api_key are correctly set in config.yml
# Test connectivity to the backend from the agent host
curl -v http://pmp-backend.example.com:8080/pmp/actuator/health

No ASH data appearing in ASH Viewer​

  • Confirm the PostgreSQL monitoring user has the pg_monitor role
  • Check that pg_stat_activity returns rows for the monitoring user:
SELECT count(*) FROM pg_stat_activity;
  • Verify the collectors.ash.enabled is set to true in config.yml

No query data in Top SQL​

  • Confirm pg_stat_statements extension is installed and enabled:
SELECT * FROM pg_stat_statements LIMIT 1;

Next Steps​