Updated k6 script with per-link metrics and file-output logging

Below is a single, ready-to-run k6 script that:

• Crawls the BASE_URL in setup() and extracts same-origin links.
• Creates per-link custom metrics (Trend for response time, Counter for requests, Counter for failures) lazily and records them for every request.
• Logs request details to console (status, URL, response time, error).
• Is configurable via environment variables: BASE_URL, REPETITIONS, DELAY_SECONDS, VUS, DURATION.
• Shows how to export metrics to a JSON file and save console logs to a file when running k6.

---

How to export metrics and logs to files

1) Export k6 metrics (including custom metrics) to JSON file

Run k6 with the --out json=metrics.json option. This writes all metrics (including the custom Trends and Counters created above) to metrics.json.

k6 run --out json=metrics.json -e BASE_URL=https://example.com -e REPETITIONS=10 -e DELAY_SECONDS=30 -e VUS=20 -e DURATION=60m load-test.js

2) Save console logs (stdout/stderr) to a file

Redirect stdout and stderr when running k6 so the console logs (the console.log / console.error lines in the script) are saved:

k6 run --out json=metrics.json -e BASE_URL=https://example.com -e REPETITIONS=10 -e DELAY_SECONDS=30 -e VUS=20 -e DURATION=60m load-test.js > k6_console.log 2>&1

• metrics.json will contain structured metric data you can parse or import into tools.
• k6_console.log will contain the per-request console lines (status, URL, response time, errors).

3) Use other outputs
k6 supports many output backends (InfluxDB, Prometheus remote write, Datadog, etc.). Replace --out json=metrics.json with the desired output (e.g., --out influxdb=http://host:8086/...) if you want to push metrics to a monitoring system.

---

How to analyze per-link metrics after the run

• metrics.json contains entries for each metric name you created (e.g., rt_example_com_path, reqs_example_com_path, fails_example_com_path). Use jq, Python, or import into a time-series DB to visualize trends and counts.
• Example quick jq to list metric names:

jq -r '.[].metric' metrics.json | sort | uniq

• For deeper analysis, import metrics.json into a script that aggregates reqs_* and fails_* per link and computes averages from rt_* Trend values.

---

Notes and recommendations

• Metric name length: metric names are sanitized and truncated; ensure uniqueness for very long paths if you need exact mapping.
• Cardinality: creating a metric per unique link increases metric cardinality. For very large sites (thousands of links) consider grouping by path prefix or sampling a subset to avoid high cardinality in your metrics backend.
• Large JSON files: --out json=metrics.json can produce large files for long runs; rotate or stream to a DB for long-term storage.
• Dynamic link discovery: for JS-rendered links, pre-generate a link list with a headless browser and feed it to k6 (e.g., via setup() reading a JSON file).
• If you want logs in structured format: instead of console logs, you can console.log(JSON.stringify({...})) and then parse the log file.