Scaling Recommendations
This guide outlines best practices for scaling dbWatch Control Center deployments in complex or growing environments.
Scaling may be necessary when:
- Managing hundreds or thousands of database instances
- Operating across multiple datacenters, cloud regions, or security zones
- Automating job operations through scripting nodes (CCC)
- Supporting multiple tenants or business units with isolated domains
Recommended Instance Hub Ratio
Each Instance Hub should typically manage up to 200–250 database instances for optimal performance and maintainability.
If managing more, consider:
- Adding additional Instance Hubs
- Distributing them by region, business unit, or technology stack
- Segmenting responsibilities (e.g., staging vs production)
In large deployments, the Domain CA (certificate authority) should run on a separate system from the Instance Hub to distribute trust responsibilities and improve security.
Recommended Cloud Router Architecture
In deployments with:
- Multiple security zones
- Hybrid on-prem/cloud environments
- Isolated customer environments (MSPs)
It is strongly recommended to:
- Deploy a Cloud Router node (or a small set of redundant routers)
- Let all Instance Hubs, CCC nodes, and Monitor Clients connect outbound only (port 7100/tcp by default) to the router(s)
- This avoids inbound firewall rules and simplifies secure interconnectivity
See:
Scale by Role Separation
Separate key responsibilities across nodes for better scalability and control.
Role | Recommendation |
Monitor Client | Deploy multiple clients near end-users or regional teams |
Instance Hub | 1 per ~200–250 instances; separate per zone if possible |
Domain CA | Place on dedicated system in larger deployments |
Dashboard / Web | Separate node or co-located with low-load hub |
CCC Script Node | Use restricted-privilege node: Scripting Node with Limited Privileges |
Efficient Instance Deployment and Job Consistency
In large environments, avoid manually installing database instances one-by-one through the GUI wizard, as this is time-consuming and error-prone.
Instead, use one of the following:
- Bulk Installing Instances
- Install many instances at once from a CSV or spreadsheet – Supported in the GUI with validation and feedback
- Registering Instance Configuration with CCC
- Automate installation from predefined XML configuration – Ideal for integration with CM tools or scripting pipelines
- Job Templates for Deployment
- Ensure consistent job configurations across environments – Centralized control over monitoring setup and updates
Job Load & Performance Tuning
- Distribute job execution across multiple hubs
- Run heavy statistics jobs off-peak
- Customize thread pool sizes in `domain-config.xml`
- Use job priorities to control resource contention
Certificate and Routing Considerations
- Each domain has its own CA hierarchy and routing privileges
- Cloud Routers validate domain-level certificates to route messages
- Avoid placing more than one node with the same key pair on the same network—see: Mirroring a Node
Monitoring at Scale
- Enable full auditing: Auditing
- Store logs on high-performance local or centralized storage
- Schedule recurring health checks or use external log collectors (e.g., SIEM)
Multi-Domain Strategy
For use cases such as MSP environments or regulatory isolation:
- Create separate domains with:
- Individual license keys