Frequently Asked Questions

Getting Started

Yes. We require a payment method on file to prevent abuse, but we won't charge you until you use resources. New accounts receive $10 credit.

What's the minimum commitment?

None. Pay-as-you-go with no contracts or minimum spend. Cancel anytime.

Can I test before committing?

Yes. $10 signup credit covers several hours of testing. Stop instances when not in use to stretch credits further.

Do you offer free tier?

Not currently. Our pricing is already competitive, and "free tier" restrictions add complexity. $10 credit covers meaningful testing.

Billing & Pricing

How does hourly billing work?

You're charged per second of usage, billed hourly. Stop an instance after 30 minutes? You pay for 30 minutes, not a full hour.

What happens when I stop an instance?

You stop paying for compute. You only pay for storage (~$0.10/GB/month). Start it again anytime - data preserved.

Are there hidden fees?

No. Transparent pricing:

Compute: per hour
Storage: per GB/month
Bandwidth: included, then $0.01/GB
No setup fees, no cancellation fees

Can I set spending limits?

Yes. Console → Account → Billing → Spending Limit. You'll receive alerts and can optionally auto-stop instances.

Do you offer discounts?

Reserved instances (30-50% off): Coming Q3 2026
Spot instances (70% off): Coming Q4 2026
Volume discounts (>$1,000/month): Contact enterprise@fugoku.com

What payment methods do you accept?

Credit/debit cards (Visa, Mastercard, Amex)
PayPal
Bank transfer ($500+/month accounts)

Technical

Can't SSH into my instance

Check firewall rules:

fugoku firewalls list-rules instance-name

Ensure port 22 is open:

fugoku firewalls add-rule instance-name \
  --protocol tcp --port 22 --source YOUR_IP/32

Verify SSH key is uploaded:

fugoku ssh-keys list

Try web console: Console → Instance → Console tab (VNC access)

Instance is slow / high CPU

Check metrics:

fugoku instances stats instance-name

If CPU maxed out:

Resize to larger plan: fugoku instances resize instance-name --plan standard-4
Investigate process causing load: top or htop

If I/O bound:

Check disk usage: df -h
Check I/O wait: iostat -x 1
Consider storage-optimized plan

How do I backup my instance?

Automated snapshots:

fugoku snapshots enable \
  --instance instance-name \
  --schedule daily \
  --retention 7

Manual snapshot:

fugoku snapshots create instance-name --name backup-2024-02-25

Offsite backup: Export critical data to separate location (AWS S3, Google Cloud Storage, etc.)

Can I migrate from AWS/GCP/Azure?

Yes. Common approaches:

1. Snapshot and transfer:

Create snapshot/image on current provider
Convert to raw/qcow2 format
Upload to Fugoku (contact support for large transfers)

2. Rsync migration:

# From old server to new Fugoku instance
rsync -avz /data/ root@fugoku-instance:/data/

3. Application-level:

Deploy application on Fugoku
Migrate database with pg_dump/mysqldump
Update DNS to point to Fugoku

Need help? Email support@fugoku.com - migrations are free.

Do you support IPv6?

Coming Q4 2026. Currently IPv4 only.

Can I use my own IP addresses?

No. We assign from our pool. Floating IPs are static and stay with your account.

What operating systems are supported?

Linux:

Ubuntu 22.04, 24.04
Debian 11, 12
CentOS / Rocky / AlmaLinux 8, 9
Fedora 39, 40
Alpine 3.18, 3.19
Arch Linux

Windows: Coming Q3 2026

Custom images: Upload your own (contact support)

Can I run Docker?

Yes. Install Docker on any instance:

curl -fsSL https://get.docker.com | sh

For GPU containers, install nvidia-docker2.

Do you support Kubernetes?

Yes. Deploy your own:

kubeadm
k3s
RKE2

Managed Kubernetes: Coming Q4 2026

Networking

Do instances have public IPs by default?

Yes. Every instance gets a public IPv4 (ephemeral). Use floating IPs for static addresses.

Can I create a private-only instance?

Yes. Don't assign public IP (option during creation). Access via VPN or bastion host.

How does private networking work?

Create VLAN with private subnet (10.0.0.0/8, 172.16.0.0/12, 192.168.0.0/16). Attach instances to network - they get second interface (eth1) with private IP.

Instances on same private network can communicate without going through internet.

Do I pay for bandwidth between instances?

No. Internal traffic (same region) is free and unlimited.

What's a floating IP?

Static public IP that can move between instances. Used for:

DNS stability (point domains at floating IP)
Failover (reassign to backup instance)
Blue-green deployments

How do load balancers work?

Layer 4 (TCP) or Layer 7 (HTTP/HTTPS) load balancing across multiple instances. Health checks automatically remove failed backends.

GPU & AI/ML

What GPUs do you offer?

NVIDIA A100 (40GB): $2.50/hour
NVIDIA H100 (80GB): $4.00/hour

Multi-GPU configurations available (2x, 4x, 8x).

Are GPUs shared or dedicated?

Dedicated. No multi-tenancy. Full GPU performance.

Can I install custom CUDA versions?

Yes. Images come with CUDA 12.x, but you can install other versions:

# Example: Install CUDA 11.8
wget https://developer.download.nvidia.com/compute/cuda/11.8.0/local_installers/cuda_11.8.0_520.61.05_linux.run
sudo sh cuda_11.8.0_520.61.05_linux.run

Do GPU instances include ML frameworks?

Yes. Pre-built images with PyTorch, TensorFlow, JAX + Jupyter Lab. Or use blank CUDA image for custom setup.

Can I run inference on CPU instances?

Yes, but slow. Use GPU for real-time inference. CPU okay for:

Small models (<100M parameters)
Batch inference (not time-sensitive)
Preprocessing / postprocessing

How do I deploy models?

See Model Deployment guide.

Quick options:

FastAPI + uvicorn
TorchServe
TensorFlow Serving
ONNX Runtime

lagos-1: Lagos, Nigeria
london-1: London, UK
frankfurt-1: Frankfurt, Germany

No cross-region data movement unless you explicitly configure it.

Current provider
Services used
Rough timeline

Cloudflare
Fastly
AWS CloudFront
BunnyCDN

Can I run containers?

Yes. Install Docker/Podman on any instance. Managed container service coming Q4 2026.

Do you support auto-scaling?

Manual scaling now. Auto-scaling coming Q3 2026. Currently:

# Add instance to load balancer
fugoku lb add-backend lb-name --instance new-instance

# Remove when load drops
fugoku lb remove-backend lb-name --instance old-instance

Can I schedule instance start/stop?

Use cron:

# Stop at 6 PM
0 18 * * * fugoku instances stop dev-instance

# Start at 8 AM
0 8 * * * fugoku instances start dev-instance

Or use a Lambda-like service (AWS Lambda, Google Cloud Functions) to call Fugoku API.

Comparison

How are you different from AWS/GCP/Azure?

Simpler:

No complex IAM roles
Transparent pricing
Actual documentation

Cheaper:

No egress fees between regions
Reserved instance pricing without commitment
Volume discounts at lower tiers

Support:

Talk to real engineers
No tier-1 scripts
<2 hour response for critical issues

How are you different from DigitalOcean/Linode?

More powerful:

GPU compute (H100/A100)
ML-optimized images
Better network performance

Better support:

Engineering team support
No outsourced tier-1
Proactive monitoring

Location:

African presence (Lagos)
European privacy (Frankfurt, London)

How are you different from Vultr/Hetzner?

Specialization:

Purpose-built for AI/ML
Pre-configured ML environments
GPU-first approach

Support:

Direct access to engineers
Not just ticket systems
Consulting available

fugoku create instance --name web-1 --tag project:ecommerce

All resources deleted
Final invoice generated
Data permanently deleted after 7 days
Email confirmation

Can't recover after 7 days.

Can I get a refund?

Eligible:

Service outage (prorated credit)
Billing errors
Unused promotional credits

Not eligible:

Usage-based charges for resources used
Third-party software licenses

Email: billing@fugoku.com

Still Have Questions?

Email: support@fugoku.com

Discord: discord.gg/fugoku

Documentation: docs.fugoku.com

Status Page: status.fugoku.com

Can't find your answer? Email support@fugoku.com - we'll add it to this FAQ.

FAQ

On this page