JSON Output Schema

lmn can write a machine-readable JSON report in addition to (or instead of) the default ASCII table output.

# Write JSON to a file (table still prints to stdout)
lmn run -f lmn.yaml --output-file result.json

# Write JSON to stdout only (no table)
lmn run -f lmn.yaml --output json

The schema is versioned. The current version is 1.

---

Top-level object

Field	Type	Always present	Description
version	integer	Yes	Schema version. Currently 1.
run	object	Yes	Execution mode and timing metadata.
requests	object	Yes	Aggregated request counts and derived metrics.
latency	object	Yes	Latency percentiles and summary statistics. All values in milliseconds.
status_codes	object	Yes	HTTP status code counts. Keys are string codes ("200", "404"). The key "error" covers connection errors with no HTTP response.
sampling	object	Yes	Reservoir sampling state at end of run.
response_stats	object | null	No	Response body field analysis. Present only when --response-template was used.
curve_stages	array | null	No	Per-stage breakdown. Present only when mode == "curve".
thresholds	object | null	No	Threshold evaluation results. Present only when thresholds were configured.

---

run

Field	Type	Description
mode	string	"fixed" or "curve"
elapsed_ms	float	Total wall-clock run time in milliseconds
curve_duration_ms	float | null	Total curve duration in milliseconds. null in fixed mode.
template_generation_ms	float | null	Time spent pre-generating request bodies. null when no template was used.

---

requests

Field	Type	Description
total	integer	Total requests sent
ok	integer	Requests with a 2xx response
failed	integer	Requests with a non-2xx response or connection error
error_rate	float	failed / total. 0.0 when total == 0
throughput_rps	float	total / elapsed_seconds. 0.0 when elapsed is zero

---

latency

All values are in milliseconds (f64). Field names carry the _ms suffix.

Field	Description
min_ms	Minimum observed latency
p10_ms	10th percentile
p25_ms	25th percentile
p50_ms	50th percentile (median)
p75_ms	75th percentile
p90_ms	90th percentile
p95_ms	95th percentile
p99_ms	99th percentile
max_ms	Maximum observed latency
avg_ms	Mean latency

Sampling

When sampling.sampled == true, latency percentiles are computed from a random sample of the full request population and are approximate. See sampling.

---

sampling

Field	Type	Description
sampled	boolean	true if latency percentiles are based on a sample rather than the full population
final_sample_rate	float	Sample rate at end of run. 1.0 means no sampling occurred.
min_sample_rate	float	Lowest sample rate observed during the run
reservoir_size	integer	Maximum results held in memory (configurable via --result-buffer)
results_collected	integer	Actual results collected (≤ reservoir_size)

---

curve_stages

Present only when mode == "curve". An array of stage objects in order.

Each stage object:

Field	Type	Description
index	integer	0-based stage index
duration_ms	float	Configured stage duration in milliseconds
target_vus	integer	Configured target VU count for this stage
ramp	string	"linear" or "step"
requests	integer	Requests sent during this stage
ok	integer	2xx responses in this stage
failed	integer	Non-2xx or error responses in this stage
error_rate	float	failed / requests for this stage
throughput_rps	float	Requests per second within this stage's duration window
latency	object	Same structure as the top-level latency object, scoped to this stage

---

thresholds

Present only when thresholds were configured. Contains aggregate counts and per-rule results.

Field	Type	Description
total	integer	Total number of threshold rules evaluated
passed	integer	Number of rules that passed
failed	integer	Number of rules that failed
results	array	Per-rule evaluation results (see below)

Each item in results:

Field	Type	Description
threshold.metric	string	Metric name (e.g. "latency_p99", "error_rate")
threshold.operator	string	Operator ("lt", "lte", "gt", "gte", "eq")
threshold.value	float	Configured threshold value
actual	float	Observed metric value
passed	boolean	Whether actual satisfied the threshold

---

response_stats

Present only when --response-template was used.

Field	Type	Description
responses_parsed	integer	Number of responses successfully matched against the template
string_fields	object	Distribution of string-valued field extractions. Outer key: field path. Inner key: extracted value. Value: count.
float_fields	object	Summary statistics for float-valued field extractions. Key: field path.
mismatch_counts	object	Count of responses where a tracked field could not be extracted. Key: field path.

Each entry in float_fields:

Field	Description
min	Minimum observed value
avg	Mean
p50	Median
p95	95th percentile
p99	99th percentile
max	Maximum observed value

---

Full example

{
  "version": 1,
  "run": {
    "mode": "fixed",
    "elapsed_ms": 2074.3,
    "curve_duration_ms": null,
    "template_generation_ms": null
  },
  "requests": {
    "total": 100,
    "ok": 100,
    "failed": 0,
    "error_rate": 0.0,
    "throughput_rps": 48.3
  },
  "latency": {
    "min_ms": 142.0,
    "p10_ms": 161.3,
    "p25_ms": 178.8,
    "p50_ms": 198.0,
    "p75_ms": 234.5,
    "p90_ms": 289.2,
    "p95_ms": 312.0,
    "p99_ms": 401.0,
    "max_ms": 487.3,
    "avg_ms": 203.1
  },
  "status_codes": {
    "200": 100
  },
  "sampling": {
    "sampled": false,
    "final_sample_rate": 1.0,
    "min_sample_rate": 1.0,
    "reservoir_size": 100000,
    "results_collected": 100
  },
  "response_stats": null,
  "curve_stages": null,
  "thresholds": {
    "total": 2,
    "passed": 2,
    "failed": 0,
    "results": [
      {
        "threshold": { "metric": "latency_p99", "operator": "lt", "value": 500.0 },
        "actual": 401.0,
        "passed": true
      },
      {
        "threshold": { "metric": "error_rate", "operator": "lt", "value": 0.01 },
        "actual": 0.0,
        "passed": true
      }
    ]
  }
}