Get started
What server capacity does your site need?
Size for tags, rates, retention, concurrency and availability, not equipment count alone.
The approach
How does it work?
- 01 Choose the scope
Choose products by the capabilities needed
- 02 Size the deployment
Size capacity and servers for the site
- 03 Deployment & support
Make included and additional work clear
- 04 Agree on a plan
An actionable initial deployment plan
Details
Scope and considerations
Reference configuration per server node
Recommended server node specifications are listed below. Final purchasing specifications are confirmed after reviewing collection volume, retention and query requirements.
| Component | Reference |
|---|---|
| CPU | 16 cores / 32 threads |
| RAM | 128 GB |
| Boot | 2 × 1 TB M.2 NVMe, OS RAID 1 |
| Hot storage | 2 SSDs, NVMe possible |
| Cold storage | 2 HDDs |
| Network | 10 GbE |
Data-disk capacity depends on retention. Additional disks can support storage and distributed processing; keep Ceph OSD server node disk counts and capacities consistent.
Evaluation and production clusters
| Configuration | Server nodes | Purpose |
|---|---|---|
| PoC | 1 | Evaluation, no HA |
| Basic HA | 3 | Smaller single site |
| Standard HA | 4 | Operational peak headroom |
| Expanded HA | 6 | Mid-sized site |
| High availability | 9 | Large single site |
| Enterprise | 12+ | Multiple sites and workload isolation |
Millisecond collection or additional tags changes the number of machines a configuration can support, so equipment count alone does not determine the number of server nodes.
Sizing inputs for consultation
Begin with machines × tags per machine × samples per second, then consider types, change patterns, retention, replication, backups and concurrent analysis. This rough rate does not guarantee compression or latency.
For existing servers, review disks, network, virtualization and administrative access as well as CPU and memory. GPU models and local LLMs are separate workloads.
Start together
Tell us what you want to improve first.
Tell us about your equipment, data and priorities to review a path from connection to operations.