How fast does each cloud actually provision?

ProvisioningIQ runs the same provisioning tests across AWS, Azure, and Google Cloud, around the clock and in multiple regions, measuring how long each cloud takes to make a resource ready. Every run uses identical settings, ensuring a fair comparison. Lower is faster.

The four things are measured here:

V1 is a virtual machine (a standard Linux server).
C1 is a serverless container.
D1 is a managed PostgreSQL database.
K1 is a managed Kubernetes cluster.

Legend

• TTH = Time-to-Health (t_total_ms): end-to-end time until the /health probe succeeds
* READY = Time-to-Ready (t_ready_ms): time until the resource is reported ready/running by the cloud
* p50 = median (50th percentile)
* p95 = 95th percentile (tail latency)
* OK/All= successful runs / total runs
* Age = time since the most recent run in that bucket (UTC)