Understanding Resource Limits

How Resources Are Allocated

Each ClawHosters instance runs on a dedicated Hetzner Cloud VPS. The tier you choose determines how much CPU, memory, and storage your instance gets. These resources are not shared with other customers.

The OpenClaw container runs inside Docker with enforced memory limits. If the container tries to use more memory than allocated, Docker will kill and restart it.

Resources by Tier

Resource	Budget	Balanced	Pro
vCPU	2 cores	4 cores	8 cores
RAM	4 GB	8 GB	16 GB
Storage	40 GB	80 GB	160 GB

What Each Resource Affects

CPU (vCPU)

CPU determines how fast your instance processes requests. More cores help when:

• Multiple users send messages simultaneously
• The instance runs compute-heavy skills or plugins
• RAG document indexing is in progress (if embedding add-on is active)

For most single-user or small-team setups, 2 cores (Budget) is enough. If you notice slow responses under load, upgrading to Balanced or Pro gives more headroom.

Memory (RAM)

Memory affects how much data the instance can hold in active memory at once. This includes:

• The Node.js runtime and OpenClaw process
• Active conversation contexts
• Loaded skills and plugins
• Temporary data from document processing

If the OpenClaw process uses too much memory, the system may restart the container automatically. You will see a brief downtime (usually under 30 seconds) when this happens.

Storage

Storage covers everything stored on disk:

• The OpenClaw Docker image itself (~2 GB)
• Conversation history and logs
• Uploaded documents and knowledge base files
• Installed packages (if you SSH in and install additional tools)
• Backups stored locally before upload (if backup add-on is active)

Storage is persistent. It survives container restarts and reboots. It is only wiped during a rebuild.

Choosing the Right Tier

Use Case	Recommended Tier
Personal assistant, single user	Budget
Small team (2-5 users), moderate load	Balanced
Production deployment, many concurrent users	Pro
RAG with large document collections	Balanced or Pro
Running multiple skills simultaneously	Pro

Start with Budget if you are unsure. You can upgrade your tier later without losing data.

What Happens When You Hit Limits

Memory Limit Exceeded

If the container uses too much memory:

1. The system terminates the container process
2. The container restarts automatically
3. The instance is unavailable for 10-30 seconds during restart
4. Active conversations in progress may lose their last message

Check if this is happening by looking at your instance's monitoring page. Frequent restarts usually mean you need a higher tier.

Storage Full

If the disk fills up:

• New messages may fail to save
• Document uploads will be rejected
• Logs stop writing, making debugging harder
• The instance may become unresponsive

Clean up unused files via SSH, or upgrade to a tier with more storage.

CPU Saturated

When all CPU cores are busy:

• Response times increase
• Messages queue up instead of processing immediately
• The instance stays functional but slow

This is usually temporary during peak usage. If it happens regularly, upgrade to a tier with more cores.

Monitoring Your Usage

The instance monitoring page in your dashboard shows current resource usage:

• CPU usage percentage
• Memory usage (current vs limit)
• Storage used vs available

Check these metrics periodically to see if your tier fits your workload. If any resource consistently runs above 80%, consider upgrading.

Upgrading and Downgrading

Tier changes are only available for monthly subscription instances. Daily-billed instances cannot change tiers. If you need a different tier on daily billing, create a new instance on the desired tier and delete the old one.

For monthly subscriptions, the change takes effect at the start of the next billing period. Upgrading provisions a new server with more resources and migrates your data. Downgrading works the same way but with fewer resources. Neither operation loses your data.

Related Documentation

• Instance Overview. General instance management
• Instance Monitoring. Checking resource usage
• Choosing Your Tier. Tier comparison and pricing
• Start, Stop, and Restart. Managing instance state