---
title: "Run the load test and read results | Grafana Labs"
description: "Run your baseline test and interpret the k6 summary output, including http_req_duration percentiles, error rates, and iteration counts."
---

# Run the load test and read results

With your load profile and checks in place, you’re ready to run the test and interpret the output. The k6 summary is dense with metrics, but for establishing a baseline you need to focus on a few key lines: `http_req_duration` for latency, `http_req_failed` for error rate, and `checks` for correctness.

To run your test and read the results, complete the following steps:

1. Open your terminal and run the test:
   
   Bash ![Copy code to clipboard](/media/images/icons/icon-copy-small-2.svg) Copy
   
   ```bash
   k6 run baseline.js
   ```
   
   k6 displays real-time progress as the test executes, showing the current VU count and completed iterations.
2. When the test completes, review the summary output. Focus on the following metrics:
   
   **`http_req_duration`** : The time it takes for requests to complete. The summary shows several aggregations:
   
   text ![Copy code to clipboard](/media/images/icons/icon-copy-small-2.svg) Copy
   
   ```text
   http_req_duration..: avg=180ms  min=95ms
                        med=165ms  max=850ms
                        p(90)=280ms  p(95)=350ms
   ```
   
   - `avg` is the mean across all requests, but averages hide outliers.
   - `p(90)` and `p(95)` are more useful for baselines. The `p(95)` value means 95% of requests completed within that time. The remaining 5% were slower.
   - If `max` is dramatically higher than `p(95)`, a small number of requests are experiencing significant delays.
   
   **`http_req_failed`** : The percentage of requests that returned a non-2xx status code:
   
   text ![Copy code to clipboard](/media/images/icons/icon-copy-small-2.svg) Copy
   
   ```text
   http_req_failed................: 0.00%   0 out of 1200
   ```
   
   The output shows how many requests failed out of the total. For a healthy baseline, this value should be at or near 0%.
   
   **`checks`** : The pass rate across all your check assertions. The checks section appears at the top of the summary under **TOTAL RESULTS**:
   
   text ![Copy code to clipboard](/media/images/icons/icon-copy-small-2.svg) Copy
   
   ```text
   checks_total.......: 2400    10.00/s
   checks_succeeded...: 100.00% 2400 out of 2400
   checks_failed......: 0.00%   0 out of 2400
   
   ✓ status is 200
   ✓ response body is not empty
   ```
   
   A 100% `checks_succeeded` rate means every response met your correctness criteria. The individual check names are listed below with ✓ or ✗ indicators.
   
   **`iterations`** : The total number of times the default function executed:
   
   text ![Copy code to clipboard](/media/images/icons/icon-copy-small-2.svg) Copy
   
   ```text
   iterations.....................: 1200    10.00/s
   ```
   
   The per-second rate represents your throughput (requests per second) during the test.
3. Confirm that your test ran cleanly: checks are passing, the error rate is near zero, and the latency values are stable. If something looks off, such as high error rates, check failures, or erratic latency, investigate before using these numbers as your baseline.
   
   > Note
   > 
   > Percentiles are more reliable than averages for baselines. An average of 200 ms could mean most requests took 150 ms but a few took 2 s. The p95 value gives you a much clearer picture of actual user experience.

In the next milestone, you identify the specific values from these results to record as your baseline.