Evaluation prerequisites and plan
Evaluate Speedscale with one service to assess replay accuracy and performance in your environment. Agree on scope and success criteria with your account team before kickoff.
Before kickoff
- Service and traffic: choose the service, endpoints and dependencies. Prepare representative requests, including authenticated flows and error cases. Test or UAT traffic is sufficient to start.
- Test environment: provide a non-production deployment, test credentials, seed data and capacity for the agreed load. Identify which dependencies will be mocked or remain live.
- Owners: name an application owner and a platform contact for installation and network access.
- Data handling: agree on fields to redact, storage location, access and retention before capture. See Data Loss Prevention.
- Access: arrange Speedscale access, installation permissions and the required binaries or container images.
Installation requirements
| Environment | Prepare |
|---|---|
| Kubernetes, classic operator | Configure an API key and follow the Quick Start. Approve RBAC and webhook access and confirm eBPF compatibility, or agree on sidecar capture. |
| Kubernetes, namespace-only | For Deployment or StatefulSet targets, follow the namespaced guide. Pre-provision the namespace and required Secrets, including the API key. Approve sidecar Pod Security exceptions and, on OpenShift, an SCC grant. eBPF is unavailable. |
| Local service | Install proxymock and activate through browser sign-in. Use an API key for headless runs. Confirm proxy routing and TLS trust; Kubernetes is optional. |
| VM or Docker | Configure an API key and review the VM or Docker guide. |
Confirm network access for your deployment, including outbound cloud connections and TCP 9443 webhook access for the classic operator. If captured traffic must stay in customer-controlled storage, review BYOC. It requires Enterprise enablement and still needs Speedscale API connectivity.
Success criteria
Agree on response-match accuracy, request rate or concurrency, test duration, error limits and p95/p99 latency targets. Confirm capture coverage, data redaction and which dependencies must be isolated. Assess results in Cloud replay reports or local proxymock replay results, and verify detection of one controlled regression.
Load tests with mocked dependencies measure the service under controlled conditions, rather than the capacity of live downstream systems.
Sample evaluation plan
| Stage | Outcome |
|---|---|
| Week 1: install and capture | Verify installation, representative traffic coverage and agreed data redaction. |
| Week 2: establish a baseline | Create a reusable recording or snapshot, configure dependency mocks and obtain a baseline replay that meets the agreed response comparisons. |
| Week 3: test and review | Run the agreed load, compare latency and errors, demonstrate regression detection and review results together. |
Agree on dates once approvals, traffic and the application are ready. Share this checklist with your account team, or contact support@speedscale.com.