Hardware Sizing Matrix
Artifactory sizing: 4 cores/6 GB (0–20 clients), 6 cores/12 GB (21–100), 8 cores/18 GB (101–200). Xray scales by indexed artifacts.
Artifactory Hardware Sizing
JFrog Artifactory sizing depends mainly on the number of active clients.
Use the following matrix to select a baseline profile, then adjust for your repository size, package types, and concurrency pattern.
| Number of Active Clients | Processor (CPU) | Memory |
|---|---|---|
| 0–20 | 4 core | 6 GB |
| 21–100 | 6 core | 12 GB |
| 101–200 | 8 core | 18 GB |
| 201+ | Contact JFrog Support | Contact JFrog Support |
Xray Hardware Sizing
JFrog Xray sizing requirements are determined by the scale of your environment.
Use indexed artifact count as the primary decision input.
| Indexed Artifacts | Xray Nodes | CPU (per node) | Memory (per node) | Xray Disk | Database | RabbitMQ |
|---|---|---|---|---|---|---|
| Up to 100,000 | 1 | 6 | 24 GB | 500 GB SSD (3000 IOPS) | 500 GB SSD (3000 IOPS) | 1 node: 4 CPU, 8 GB, 100 GB SSD |
| Up to 1,000,000 | 2 | 4 | 8 GB | 300 GB per node | 8 CPU, 32 GB, 500 GB SSD (3000 IOPS) | 3 nodes (split): 4 CPU, 8 GB, 100 GB SSD each |
| Up to 2,000,000 | 3 | 6 | 12 GB | 300 GB per node | 16 CPU, 32 GB, 1 TB SSD (3000 IOPS) | 3 nodes (split): 4 CPU, 8 GB, 100 GB SSD each |
| Up to 10,000,000 | 3 | 8 | 24 GB | 300 GB per node | 16 CPU, 64 GB, 2.5 TB SSD (3000 IOPS) | 3 nodes (split): 4 CPU, 8 GB, 100 GB SSD each |
| Over 10,000,000 | Contact JFrog Support | Contact JFrog Support | Contact JFrog Support | Contact JFrog Support | Contact JFrog Support | Contact JFrog Support |
Non-Kubernetes Deployment
Use this section for VM or bare-metal deployments.
Up to 100K indexed artifacts – No High Availability
| Component | Nodes | CPU Cores | Memory | Disk Space |
|---|---|---|---|---|
| Xray and DB | 1 | 6 | 24 GB | 500 GB SSD (3000 IOPS) |
| RabbitMQ | 1 | 4 | 8 GB | 100 GB SSD (3000 IOPS) |
| JFrog Advanced Security | 1 | 6 | 24 GB | 500 GB SSD (3000 IOPS) |
Up to 1M indexed artifacts
| Component | Nodes | CPU Cores | Memory | Disk Space |
|---|---|---|---|---|
| Xray | 2 | 4 | 8 GB | 300 GB |
| DB | 1 | 8 | 32 GB | 500 GB SSD (3000 IOPS) |
| RabbitMQ (Split) | 3 | 4 | 8 GB | 100 GB SSD (3000 IOPS) each |
| JFrog Advanced Security | 2 | 8 | 24 GB | 300 GB |
Up to 2M indexed artifacts
| Component | Nodes | CPU Cores | Memory | Disk Space |
|---|---|---|---|---|
| Xray | 3 | 6 | 12 GB | 300 GB |
| DB | 1 | 16 | 32 GB | 1 TB SSD (3000 IOPS) |
| RabbitMQ (Split) | 3 | 4 | 8 GB | 100 GB SSD (3000 IOPS) each |
| JFrog Advanced Security | 4 | 8 | 24 GB | 300 GB |
Up to 10M indexed artifacts
| Component | Nodes | CPU Cores | Memory | Disk Space |
|---|---|---|---|---|
| Xray | 3 | 8 | 24 GB | 300 GB |
| DB | 1 | 16 | 64 GB | 2.5 TB SSD (3000 IOPS) |
| RabbitMQ (Split) | 3 | 4 | 8 GB | 100 GB SSD (3000 IOPS) each |
| JFrog Advanced Security | 8 | 8 | 24 GB | 300 GB |
Over 10M indexed artifacts: Contact JFrog Support
Related Topics
- Requirements Matrix
- Software Dependencies Matrix
- Xray and RabbitMQ Nodes Recommendations
- General System Requirements
- Install JFrog Platform via Helm Quick Start
Frequently Asked Questions
This section provides answers to frequently asked questions.
FAQs
Q: What is the primary factor for sizing JFrog Artifactory hardware?
A: Artifactory sizing depends mainly on the number of active clients. The baseline starts at 4 cores and 6 GB RAM for up to 20 active clients, scaling to 8 cores and 18 GB RAM for 101–200 clients. Deployments with more than 200 active clients require guidance from JFrog Support.
Q: How many Xray nodes are required to index up to 1 million artifacts?
A: Indexing up to 1 million artifacts requires 2 Xray nodes, each with 4 CPU cores and 8 GB of RAM, with 300 GB of disk per node. The database should be a separate node with 8 CPU, 32 GB RAM, and a 500 GB SSD at 3000 IOPS.
Q: When does RabbitMQ need to run on separate servers from Xray?
A: For Xray HA deployments — meaning more than one Xray node or more than 100,000 indexed artifacts — RabbitMQ and Xray must run on separate servers (split mode). Single-node, non-HA RabbitMQ remains supported for smaller deployments.
Q: How many RabbitMQ nodes are recommended for a Xray Kubernetes cluster, and why?
A: JFrog recommends a three-node RabbitMQ cluster for Xray. Quorum Queues require at least three nodes for one-node fault tolerance, and JFrog applications are tuned for three-node configurations. Larger clusters may lower throughput due to replication overhead.
Q: What are the minimum hardware requirements for JFrog Workers?
A: JFrog Workers requires a minimum of 4 CPU cores, 4 GB of RAM, and 50 GB SSD. The minimum Artifactory version required to run Workers is 7.98.4.
RabbitMQ Note
RabbitMQ is essential for Xray and acts as a message broker between application services.
See: Xray and RabbitMQ Nodes Recommendations
For Xray HA (more than one node or more than 100K indexed artifacts), RabbitMQ and Xray should run on separate servers (split mode).
See: RabbitMQ Split for Xray
Kubernetes Deployment
YAML configuration files for sizing Artifactory, Xray, and Distribution are available on our GitHub.
RabbitMQ Quorum Queues Support (Helm Chart ≥ 103.124)
- JFrog Xray supports RabbitMQ Quorum Queues starting from Helm chart version 103.124.
- Quorum Queues require at least three RabbitMQ nodes for one-node fault tolerance.
- Single-node (non-HA) RabbitMQ remains supported.
- Recommendation: Use a three-node RabbitMQ cluster for Xray. Larger clusters may lower throughput due to replication overhead. JFrog apps are tuned for three-node configurations.
Distribution Hardware Sizing
Minimum requirements when running Distribution with an external database. Actual requirements may vary with your data size.
| Processor | Memory | Disk Space |
|---|---|---|
| 4 cores | 8 GB | 200 GB SSD |
Workers Hardware Sizing
Minimum requirements when running Workers:
| Processor | Memory | Disk Space |
|---|---|---|
| 4 cores | 4 GB | 50 GB SSD |
Minimum Artifactory Version
The minimum Artifactory version needed to run Worker is 7.98.4 and above.
AppTrust Hardware Sizing
Minimum requirements when running AppTrust:
| Service | Processor | Memory | Disk Space |
|---|---|---|---|
| AppTrust | 2 Cores | 1 GB | 50 GB SSD |
| UnifiedPolicy | 2 Cores | 1 GB | 50 GB SSD |
Frontend, JFBus, and JFmelt Hardware Sizing
Frontend, JFBus, and JFmelt run as standalone Kubernetes Deployments by default (see Service Pod Architecture Changes). These are stateless services with no dedicated disk requirement; size CPU and memory per your Artifactory sizing profile (Helm chart sizing/ values).
Frontend — resource requirements scale with your Artifactory sizing profile:
| Sizing Profile | CPU Request | Memory Request | Memory Limit |
|---|---|---|---|
| Small | 100m | 250Mi | 500Mi |
| Medium | 100m | 500Mi | 1Gi |
| Large | 200m | 800Mi | 2Gi |
| XLarge | 1 core | 1Gi | 2Gi |
JFBus and JFmelt — the same values apply across all sizing profiles (Small through XLarge):
| Service | CPU Request | Memory Request | Memory Limit |
|---|---|---|---|
| JFBus | 500m | 3500Mi | 3500Mi |
| JFmelt | 500m | 500Mi | 1Gi |
Updated 12 days ago