Skip to main content

PerForge v2.1.15: Gatling Support, Custom Filters, and Grafana 12

· 3 min read
Uladzislau Shklianik
Uladzislau Shklianik
Performance engineer
Siamion Viatoshkin
Siamion Viatoshkin
Performance engineer

PerForge v2.1.15 brings Gatling into the fold, gives you more control over which tests appear in the test list, and keeps pace with the latest Grafana release. 🎉

1. Gatling InfluxDB v2 support (BETA)​

Gatling users can now connect PerForge to their test results stored in InfluxDB v2. Since Gatling publishes metrics over the Graphite protocol rather than writing directly to InfluxDB, the recommended pipeline is:

Gatling → Graphite (TCP :2003) → Telegraf → InfluxDB v2

Telegraf acts as the bridge and requires a template that maps the Graphite path segments to proper InfluxDB tags:

[[inputs.graphite]]
  listen = ":2003"
  protocol = "tcp"
  templates = [
    "gatling.*.*.*.*.* measurement.testTitle.simulation.request.status.field"
  ]

[[outputs.influxdb_v2]]
  urls = ["http://your-influxdb-host:8086"]
  token = "your-api-token"
  org = "your-org"
  bucket = "gatling"

Each Gatling run should use a unique rootPathPrefix so runs are stored separately:

TEST_RUN_ID="run_$(date +%Y%m%d_%H%M%S)"
./gatling.sh -Dgatling.data.graphite.rootPathPrefix=gatling.${TEST_RUN_ID}

Once data is in InfluxDB, configure the integration in PerForge by selecting gatling_influxdb_v2 (BETA) as the listener. All standard PerForge features work out of the box: test list, aggregated table, RPS/response time/error charts, ML analysis, and report generation.

A few Gatling-specific notes:

  • Gatling does not emit a 90th percentile — PerForge uses percentiles95 instead, and all labels in the UI and reports automatically reflect this.
  • InfluxDB 1.8 is not supported for Gatling; only InfluxDB v2 is available.

Full setup details, including custom variables and multi-node runs, are covered in the Integrations documentation.

2. Custom filter tags​

InfluxDB integrations now support custom filter tags — additional tag conditions used when fetching the list of test titles on the tests page. This is useful when a single bucket holds data from multiple tools, environments, or simulations and you only want certain runs to appear in the test list.

Only test titles that match the filter will be shown in the tests tab. The filter is configured per integration in the InfluxDB settings form.

3. Start time offset​

Load tests typically begin with a ramp-up phase where virtual users are gradually added. Metrics collected during this period are often unstable and can skew aggregated results and analysis.

The new Start time offset field (in minutes) shifts the detected test start time forward by the specified amount, effectively trimming the ramp-up from all calculations. Set it to the duration of your ramp-up phase and PerForge will base aggregated data, charts, and ML analysis only on the stable portion of the test.

4. Grafana 12 support​

PerForge now works with Grafana 12, which introduced string-based panel IDs alongside the older integer IDs. Dashboard rendering used in reports picks up the correct format automatically — no configuration changes required on your side.

The dashboardScene feature flag introduced in Grafana 12 is also handled, ensuring that screenshot rendering continues to work reliably across all supported Grafana versions.

We hope these additions make PerForge more flexible for a wider range of performance testing setups. Happy perforging! 🔥

New version