Choose a PMM deployment strategy

Whether you’re monitoring a single database or managing hundreds across your organization, it’s important to select the appropriate deployment approach for both PMM Server and PMM Client components.

Plan your PMM architecture to align with your infrastructure, growth expectations, and operational needs.

PMM architecture overview

PMM can be deployed in flexible ways depending on your infrastructure and monitoring needs. Its architecture consists of two main components:

• PMM Server: The central component that stores, analyzes, and visualizes monitoring data
• PMM Client: The distributed component installed on database hosts to collect metrics

Planning considerations

Hardware and network requirements

For detailed hardware and network specifications, see:

• Hardware and system requirements
• Network and firewall requirements

Architecture considerations

• Consider network segmentation and access controls
• Plan user authentication and authorization strategy
• Evaluate TLS certificate requirements (self-signed vs custom certificates)
• For high-security environments, consider Podman’s rootless container capabilities
• Both binary installation and Docker containers can be run without root privileges, enhancing security

For information on PMM’s architecture, see PMM architecture.

PMM Server deployment options

Method	Best for	Advantages	Considerations
Docker	Development, testing & production	✔ Quick setup ✔ Simple upgrades ✔ Works in various environments	⚠ Requires Docker knowledge ⚠ May need additional configuration for production
Podman	Security-focused setups	✔ Rootless containers ✔ Enhanced security ✔ OCI-compatible	⚠ Requires Podman installation & knowledge
Helm	Cloud-native environments	✔ Scalable & high availability ✔ Kubernetes-native	⚠ Requires existing Kubernetes cluster ⚠ More complex setup
Virtual Appliance	Traditional environments	✔ Pre-configured with all dependencies ✔ Dedicated resources	⚠ Larger resource footprint ⚠ Requires a hypervisor

PMM Client deployment options

Deployment method	Best for	Advantages	Considerations
Package Manager	Standard Linux environments	• Easy install • Native to OS	• OS-specific • Requires repo access
Binary Package	Custom/isolated environments	• Portable • Minimal dependencies	• Manual install & updates
Docker	Containerized hosts	• Consistent environment • Easy to manage	• Requires Docker • Needs access to host metrics

Recommended deployment patterns

Based on the scale and environment of your monitoring needs, we recommend different deployment patterns:

Small-scale (1-30 database instances)Medium (31-200)Large (200+)Cloud-based database monitoring

• PMM Server: Docker or Virtual Appliance
• PMM Client: Package Manager
• Implementation tips:
  • for Docker, use the easy install script for quick setup
  • for Virtual Appliance, use the pre-configured OVA file
  • consider backup options early, even for small deployments
• Ideal for: Small businesses, development environments, initial deployments

• PMM Server: Docker with volume storage or Kubernetes
• PMM Client: Package Manager or Docker
• Implementation tips:
  • use Docker volumes instead of host directories for better data management
  • consider setting up high availability for production environments
  • implement regular backup procedures for monitoring data
• Ideal for: Mid-sized companies, production environments

• PMM Server: Kubernetes with proper resource allocation
• PMM Client: Automated deployment via package manager
• Implementation tips:
  • use infrastructure as code to manage deployments
  • consider distributed monitoring architecture
  • implement proper monitoring of the PMM Server itself
• Ideal for: Large enterprises, mission-critical database fleets

• PMM Server: AWS Marketplace (for AWS) or Kubernetes (for other clouds)
• PMM Client: Package Manager or automated cloud deployment
• PMM Remote: For monitoring cloud database services (RDS, Azure DB, Cloud SQL)
• Implementation tips:
  • use cloud-native storage options for better performance
  • leverage auto-scaling groups for handling variable loads
  • consider network costs when planning your architecture
• Ideal for: Cloud-native companies, hybrid cloud environments

Deployment planning checklist

Review this checklist to help you plan and size your monitoring environment and ensure your PMM environment is efficient, secure, and scalable from day one:

✓ Inventory of systems - Document all database instances that need monitoring

✓ Estimate monitoring scope - Calculate number of instances and expected metric volume

✓ Size the PMM Server - Determine hardware requirements based on monitoring load

✓ Choose Server deployment method - Select the appropriate PMM Server installation option

✓ Select Client install methods - Identify the best PMM Client setup for each system type

✓ Verify network access - Ensure proper connectivity and firewall rules are in place

✓ Plan data retention - Establish backup and disaster recovery processes

✓ Define maintenance - Create upgrade and patching procedures for PMM components

Next step

Hardware and system requirements