Upgrade metrics service

Safely upgrade metrics service when upgrading Ansible Automation Platform to benefit from new features and bug fixes while maintaining continuous data collection.

Before you begin

  • Current metrics service installation is operational
  • Backup of metrics_service database completed
  • New version of Ansible Automation Platform containerized installer downloaded

Procedure

  1. Back up current configuration 
    cp /path/to/installer/inventory /path/to/backup/inventory.backup.$(date +%Y%m%d)
pg_dump -h localhost -U metrics_service metrics_service > metrics_service_backup_$(date +%Y%m%d).sql
  2. Stop metrics service 
    systemctl --user stop automation-metrics-web.service automation-metrics-tasks.service automation-metrics-scheduler.service
  3. Run upgrade installer 
    cd /path/to/new-installer
ansible-playbook -i inventory install.yml
    Note:

    There is no separate upgrade.yml playbook. The install.yml playbook handles upgrades. The role's tasks/update.yml automatically detects image changes, stops services, and recreates containers with the new image.

  4. Verify database schema migration

    Database migrations are handled automatically during upgrade via the automation-metrics-init container, which executes Django migrations through /app/scripts/init.sh. The process retries up to 5 times if failures occur.

    # Check init container logs for migration status
podman logs automation-metrics-init

# Verify database schema
psql -h localhost -U metrics_service -d metrics_service -c "\dt"
    Note:

    The update.yml task in the automationmetrics role automatically handles container image updates and sets a recreate flag when newer images are detected.

  5. Start metrics service 
    systemctl --user start automation-metrics-web.service automation-metrics-tasks.service automation-metrics-scheduler.service

Rollback procedure

Rollback is not supported. Metrics service does not provide rollback functionality.

Recovery procedure if upgrade fails:

  1. Uninstall metrics service.
  2. Reinstall previous version using previous installer

Important limitations:

  • Backup & restore is not guaranteed to work if the version goes down (version downgrade)
  • Django migrations may not be reversible
  • Database schema changes may be incompatible with older versions

Recommendation: Test upgrades in non-production environments first to validate upgrade path and verify data integrity before upgrading production systems.