lightspeed/ops/large_data_harness
Deterministic large-data and async-runtime expansion harness for M28.
Types
Heavy-workload benchmark metrics.
pub type Benchmark {
Benchmark(
workload: String,
p50_latency_ms: Int,
p95_latency_ms: Int,
avg_payload_bytes: Int,
avg_patch_ops: Int,
peak_memory_units: Int,
throughput_events_per_second: Int,
root_churn_events: Int,
)
}
Constructors
-
Benchmark( workload: String, p50_latency_ms: Int, p95_latency_ms: Int, avg_payload_bytes: Int, avg_patch_ops: Int, peak_memory_units: Int, throughput_events_per_second: Int, root_churn_events: Int, )
Versioned M28 budget profile.
pub type Budget {
Budget(
workload: String,
max_p95_latency_ms: Int,
max_avg_payload_bytes: Int,
max_avg_patch_ops: Int,
max_peak_memory_units: Int,
min_throughput_events_per_second: Int,
max_root_churn_events: Int,
)
}
Constructors
-
Budget( workload: String, max_p95_latency_ms: Int, max_avg_payload_bytes: Int, max_avg_patch_ops: Int, max_peak_memory_units: Int, min_throughput_events_per_second: Int, max_root_churn_events: Int, )
Budget-check result.
pub type BudgetResult {
BudgetResult(workload: String, passed: Bool, reason: String)
}
Constructors
-
BudgetResult(workload: String, passed: Bool, reason: String)
Full M28 report.
pub type Report {
Report(
outcomes: List(ScenarioOutcome),
failed_scenarios: Int,
nondeterministic_failures: Int,
)
}
Constructors
-
Report( outcomes: List(ScenarioOutcome), failed_scenarios: Int, nondeterministic_failures: Int, )
M28 conformance scenarios.
pub type Scenario {
LargeListWindowedIncremental
LargeGridWindowedIncremental
LargeChartWindowedIncremental
AsyncBackpressureRecovery
HeavyBudgetEnforcement
}
Constructors
-
LargeListWindowedIncremental -
LargeGridWindowedIncremental -
LargeChartWindowedIncremental -
AsyncBackpressureRecovery -
HeavyBudgetEnforcement
Values
pub fn benchmark_signature(benchmark: Benchmark) -> String
Stable benchmark signature.
pub fn budget_failures(results: List(BudgetResult)) -> Int
Number of failing budget checks.
pub const budget_version: Int
pub fn evaluate_budget(
benchmarks: List(Benchmark),
budgets: List(Budget),
) -> List(BudgetResult)
Evaluate M28 benchmark suite against one budget profile.
pub fn nondeterministic_failures(report: Report) -> Int
Nondeterministic scenario count.
pub fn run_heavy_benchmarks() -> List(Benchmark)
Run deterministic benchmark suite for M28.
pub fn run_scenario(scenario: Scenario) -> ScenarioOutcome
Run one scenario twice and require deterministic signature parity.
pub fn snapshot_report_markdown() -> String
Deterministic markdown report for fixture scripts.
pub fn snapshot_signature() -> String
Deterministic snapshot signature for fixture drift gates.
pub const snapshot_version: Int