About

ProvisioningIQ measures cloud infrastructure provisioning latency. Continuous synthetic tests run around the clock across AWS, Azure, and GCP, provisioning real VMs, serverless containers, and managed Postgres databases, then measuring how long each step actually takes.

Why this exists

Cloud providers publish availability SLAs. No one publishes how long it takes to create a resource. That gap matters whenever you spin up CI environments, manage IaC at scale, or evaluate which cloud best fits your workload, and there’s been no publicly available comparable dataset to reference.

ProvisioningIQ fills that gap.

How it works

Tests run every few hours per region. Each test:

1. Submits a creation request to the provider’s API
2. Polls until the resource is healthy and reachable
3. Records every phase timing (API submit, ready, probe-OK, total)
4. Tears the resource down and records destroy latency

Coverage: 3 providers, 9 regions, the smallest burstable instance class on each. The metric definitions are documented in the dashboard’s Legend section.

Independence

ProvisioningIQ is built and maintained by AppsWireless. No cloud provider sponsors, funds, or contributes data to this project. Every dollar of cloud spend used to run the benchmarks is paid out of pocket.

If a finding looks bad for one provider, that’s because the data says so.