magnus919_agent-skills/capacity-and-cost-engineering/templates/load-soak-test-plan.md

Load / Soak Test Plan

Fill this template to design a load or soak test that produces valid capacity evidence. A capacity claim without load/soak evidence is incomplete.

Test identification

• Test objective: [fill: what capacity decision does this test support? e.g., "validate that the service can handle projected peak + 20% headroom at P99 latency < 200ms"]
• Test type: [fill: load test (target throughput, short duration) / soak test (sustained load, extended duration) / both]
• Test owner: [fill: name or team]

Target throughput

• Target throughput: [fill: e.g., 1000 requests/second]
• Rationale: [fill: e.g., projected peak demand 850 req/s + 20% headroom]
• Ramp-up profile: [fill: how quickly does load ramp to target? e.g., linear over 5 minutes, step function]
• Steady-state duration (load test): [fill: e.g., 30 minutes after ramp-up]
• Soak duration (if applicable): [fill: e.g., 24 hours]

Test environment

• Environment: [fill: production-like staging, dedicated test environment, or production (state which)]
• Environment parity: [fill: how does this environment differ from production? instance sizes, data volumes, network topology — every difference is an assumption]
• Justification: [fill: why is this environment sufficient evidence? if it is not production-like, what is the risk of extrapolation?]

Success criteria

Criterion	Target	Measurement method
P50 latency	[fill: e.g., < 50ms]	[fill: client-side histogram, server-side metric]
P99 latency	[fill: e.g., < 200ms]	[fill: client-side histogram]
Error rate	[fill: e.g., < 0.1% non-5xx]	[fill: response status aggregation]
CPU utilization	[fill: e.g., < 70% sustained]	[fill: OS metrics / container metrics]
Memory utilization	[fill: e.g., stable — no upward trend over soak]	[fill: OS metrics; max threshold e.g., 80%]
[fill: other]	[fill: ...]	[fill: ...]

Data collection

Metric	Source	Granularity	Retention
Latency percentiles	[fill: load generator]	[fill: 1-second buckets]	[fill: duration of test + 30 days]
Resource utilization	[fill: infrastructure metrics]	[fill: 10-second intervals]	[fill: duration of test + 30 days]
Error logs	[fill: application logs]	[fill: per-event]	[fill: duration of test + 30 days]
[fill: ...]	[fill: ...]	[fill: ...]	[fill: ...]

Soak-specific checks

For soak tests, additionally monitor and record:

• Memory trend (flat, slowly rising, or leaking — with rate estimate)
• File descriptor count (stable or growing)
• Connection pool utilization (stable or growing)
• Disk usage growth (log rotation, temp files, WAL growth)
• GC pause time and frequency (stable or degrading)
• Any metric with a statistically significant trend over the soak period

Test boundary

• Boundary exercised: [fill: component / integration / end-to-end / production]
• What this test does NOT cover: [fill: e.g., does not test cross-region failover, does not exercise all API endpoints, does not include cold-start scenarios]

Evidence record

After the test, complete this section:

• Test date: [fill: YYYY-MM-DD]
• Actual throughput achieved: [fill: ...]
• P50 latency (observed): [fill: ...]
• P99 latency (observed): [fill: ...]
• Error rate (observed): [fill: ...]
• Peak CPU utilization: [fill: ...]
• Peak memory utilization: [fill: ...]
• Soak findings (if applicable): [fill: memory trend, FD trend, connection pool trend, any anomalies]
• Verdict: [fill: PASS — all success criteria met / FAIL — criteria not met (list which) / PASS WITH GAPS — anomalies found (list which)]
• Follow-up actions: [fill: any findings requiring investigation or re-test]

Assumptions

• [fill: test environment is sufficiently representative of production]
• [fill: load profile (request mix, payload sizes, user behavior) is representative of real traffic]
• [fill: no external dependencies have different behavior in test vs production]
• [fill: any other unverified assumption]

Ownership

• Test designer: [fill: name or team]
• Test executor: [fill: name or team — may differ from designer]
• Evidence reviewer: [fill: name or role who reviews and accepts the evidence]