L — the headroom tier

Six vCPUs, eight gigabytes of RAM, 120 GB NVMe SSD, 8 TB bandwidth — for production workloads that have outgrown the 4-GB M but don’t yet need an XL.

What it’s good for

• Production SaaS for medium teams (10–50 MAUs sustained, 100s peak)
• Database hosts with working sets that need to live in RAM
• Self-hosted Coolify / Dokploy serving more than a handful of sites
• Big CI runners with parallel build queues

What it’s not good for

• 16+ GB RAM workloads (use XL)
• Dedicated bare-metal allocation (use Custom)

Hardware lineage

Same enterprise hardware, dual-ISP redundancy, hydropower. A reserved-allocation slice — you get the RAM and the vCPU shares you paid for, every time.