如何在JMeter中从JTL文件生成含多指标的易读图表？

Let me walk you through exactly how to capture and visualize those metrics (active users, error rate, detailed response times, and TPS) using your CSV-formatted JTL file.

First: Make Sure Your CSV Has All the Required Raw Data #

Your current jmeter.save.saveservice.output_format=csv setting is a good start, but it doesn’t capture all the data needed for those extra metrics. You’ll need to update your user.properties file to include these additional configurations:

# Capture error status and codes to calculate error rate
jmeter.save.saveservice.success=true
jmeter.save.saveservice.response_code=true

# Capture thread/active user data
jmeter.save.saveservice.threads=true
jmeter.save.saveservice.thread_name=true

# Keep timestamp and transaction labels for TPS calculations
jmeter.save.saveservice.timestamp=true
jmeter.save.saveservice.label=true

Restart JMeter after updating the file, then re-run your test to generate a new CSV that includes all these fields. If you don’t want to re-run the test, double-check your existing CSV—if it already has these columns, you can skip this step.

Visualizing Each Metric #

Active Users #

To see how many concurrent users were active during your test (aligned with request count on the X-axis), use JMeter’s Active Threads Over Time listener. Load your CSV into this listener, and it will plot the number of active threads (your active users) against either time or request count (you can adjust the X-axis setting in the listener’s options).

Error % (Error Rate) #

• For a quick numeric breakdown, use the Summary Report listener—it automatically calculates the error percentage per transaction and overall.
• For a visual chart, use the Errors per Second listener to track error frequency, or use the Custom Graph listener to create a series that calculates error rate (dividing failed requests by total requests) and plots it against your request count X-axis.

Response Time Details #

You already have average/median/percentile curves, but if you want more granular views:

• Use the Response Time Percentiles listener to get a clearer view of how your 90%/95% (or even 99%) percentiles behave across request count.
• The Response Times Distribution listener can show you how responses are spread across different time buckets, helping you spot outliers or patterns.

Transactions per Second (TPS) #

JMeter’s Transactions per Second listener is built for this. Load your CSV into it, and it will plot the number of completed transactions per second. If you want to align it with your "request count" X-axis instead of time, you can customize the listener’s grouping settings to aggregate by request count intervals.

Bonus: Combine Multiple Metrics in One Chart #

If you want to view TPS, error rate, and active users all in a single visualization, use the Custom Graph listener. It lets you add multiple data series from your CSV, adjust axes (e.g., use a secondary Y-axis for TPS/error rate), and match the X-axis to your request count.

内容的提问来源于stack exchange，提问作者ChanGan