You can now retrieve real-time performance data for your virtual machines. Supported metrics: cpu, memory, network and disk. Use the range parameter to query over 5m to 24h windows.

Results are cached and updated continuously.

→ See your metrics

You can now use User Data scripts when deploying virtual machines to automate initialization tasks like package installation, configuration, and bootstrapping from the moment your VM boots up.

→ Deploy now

→ Learn more about User Data

Object Storage is now available on the platform. S3-compatible, store and retrieve any amount of data using standard S3 APIs and the tools you already know.

Two storage classes. No request fees, retrieval fees, or minimum terms:

High-Performance

Low-latency, high-throughput storage for AI and general-purpose workloads

• Pricing: $0.024 / GB-month
• Availability: New York

Standard

Ideal for backups, archives, and long-term retention

• Pricing: $0.0085 / GB-month
• Availability: Ashburn and Dallas

New versions of Block Storage and File Storage are now in preview and available by request.

All three Latitude.sh storage products offered are free of egress charges.

→ Create your bucket

→ Learn more

Our first Canadian site is now available!

Get started in TOR today with our latest series of bare metal servers:

• f4.metal.small ($ 0.55/hr)
• m4.metal.large ($ 2.03/hr)
• m4.metal.xlarge ($ 4.76/hr)

More instance types are on their way and will be added to the platform.

For custom builds, just reach out to our sales team → Contact

You can now dynamically scale the number of nodes in your Kubernetes clusters directly from the dashboard, without recreating the cluster.

From the Nodes page of any ready cluster, click Scale... to open the scaling modal. You can adjust:

• Control plane nodes: Set the count from 1 up to your plan's available capacity
• Worker nodes: Scale from 0 up to the limit, or add workers to a cluster that was created without them

The modal displays per-node pricing and your total hourly cost as you adjust the counts, so you can see the impact before confirming. Control plane and worker scaling requests run in parallel for faster execution.

For clusters created without workers, select a worker plan from the dropdown before adding your first worker nodes. Once set, subsequent scaling operations use that plan automatically.