Select the right Firecracker static CPU template (C3/T2/T2A/T2CL/T2S) for your host CPU family when homogenizing a heterogeneous fleet

Documented steps

1. Identify host CPU vendor+model (lscpu on x86, or /proc/cpuinfo) — static templates are vendor-specific: Intel templates cannot run on AMD and vice versa
2. Choose a template that matches your fleet's common denominator: T2/T2C for Intel, T2A for AMD Milan, T2CL for Cascade/Ice Lake, T2S for migrating snapshots between Skylake and Cascade Lake, C3 for lowest latency
3. Apply via PUT /machine-config with the cpu_template field BEFORE starting the machine: curl --unix-socket /tmp/firecracker.socket -X PUT http://localhost/machine-config -H 'Accept: application/json' -H 'Content-Type: application/json' -d '{"vcpu_count":2,"mem_size_mib":1024,"cpu_template":"T2CL"}'
4. Verify by GET /machine-config after boot to confirm cpu_template is set on the running microVM
5. Note static CPU templates are DEPRECATED since Firecracker v1.5.0 and scheduled for removal; prefer a custom CPU template (via /cpu-config) for new deployments