How to Visualize JMeter Performance Data with Grafana, InfluxDB, and Prometheus

Performance testing often requires monitoring many components such as OS, middleware, queues, caches, databases, and front‑end services. This article focuses on the most common monitoring stack—Grafana, InfluxDB, Prometheus, and Exporters—and shows how to integrate JMeter and node_exporter into it.

1. JMeter + InfluxDB + Grafana data flow

JMeter normally displays results in its console, plugins, or generated HTML, but these methods waste time and memory. Using the Backend Listener, JMeter can asynchronously send metrics (TPS, response time, thread count, error rate) to InfluxDB every 30 seconds (configurable via #summariser.interval in jmeter.properties).

The InfluxDB schema creates two measurements: events (stores test‑tile and event tags) and jmeter (stores per‑transaction statistics). Grafana reads from the jmeter measurement to plot time‑series graphs.

Backend Listener configuration

private void addMetrics(String transaction, SamplerMetric metric) {
    // ALL status
    addMetric(transaction, metric.getTotal(), metric.getSentBytes(), metric.getReceivedBytes(), TAG_ALL, metric.getAllMean(), metric.getAllMinTime(),
            metric.getAllMaxTime(), allPercentiles.values(), metric::getAllPercentile);
    // OK status
    addMetric(transaction, metric.getSuccesses(), null, null, TAG_OK, metric.getOkMean(), metric.getOkMinTime(),
            metric.getOkMaxTime(), okPercentiles.values(), metric::getOkPercentile);
    // KO status
    addMetric(transaction, metric.getFailures(), null, null, TAG_KO, metric.getKoMean(), metric.getKoMinTime(),
            metric.getKoMaxTime(), koPercentiles.values(), metric::getKoPercentile);
    metric.getErrors().forEach((error, count) -> addErrorMetric(transaction, error.getResponseCode(),
                error.getResponseMessage(), count));
}

Metrics are then sent to InfluxDB via writeAndSendMetrics, which builds line protocol strings and posts them with an HTTP request.

2. InfluxDB storage example

> show databases
name
----
_internal
jmeter
> use jmeter
Using database jmeter
> show measurements
name
----
events
jmeter
> select * from jmeter where application='7ddemo' limit 10
...

The jmeter measurement contains fields such as avg, count, pct95.0, etc., which Grafana queries to draw TPS and response‑time curves.

3. Grafana dashboard setup

Configure an InfluxDB data source (URL, database, user, password) and import the official JMeter dashboard (ID 5496). Example queries:

SELECT last("count") / $send_interval FROM "$measurement_name" WHERE ("transaction" =~ /^$transaction$/ AND "statut" = 'ok') AND $timeFilter GROUP BY time($__interval)

for TPS, and

SELECT mean("pct95.0") FROM "$measurement_name" WHERE ("application" =~ /^$application$/) AND $timeFilter GROUP BY "transaction", time($__interval) fill(null)

for the 95th‑percentile response time.

4. node_exporter + Prometheus + Grafana data flow

node_exporter collects OS metrics (CPU, memory, etc.) from /proc and exposes them to Prometheus. After downloading and extracting Prometheus, add a job to prometheus.yml:

- job_name: 's1'
  static_configs:
  - targets: ['172.17.211.143:9200']

Start Prometheus and configure Grafana to use it as a data source. Import the official node_exporter dashboard (ID 11074).

Example Prometheus queries for CPU usage

avg(irate(node_cpu_seconds_total{instance=~"$node",mode="system"}[30m])) by (instance)
avg(irate(node_cpu_seconds_total{instance=~"$node",mode="user"}[30m])) by (instance)
avg(irate(node_cpu_seconds_total{instance=~"$node",mode="iowait"}[30m])) by (instance)
1 - avg(irate(node_cpu_seconds_total{instance=~"$node",mode="idle"}[30m])) by (instance)

These queries retrieve CPU usage percentages, which are displayed alongside top‑like values for verification.

5. Summary

Understanding where monitoring data originates—whether from JMeter’s Backend Listener, node_exporter, or other Exporters—is essential for accurate performance analysis. The described stack provides a real‑time, queryable view of test metrics and system resources, avoiding the overhead of HTML reports while preserving the ability to compare and replay data after a test run.