Test Types
Tigrister ships five built-in test types. Each one is a named preset that lives on the spec with its own independent configuration — switching between types never loses the other settings, so you can design a Load run, a Stress run and a Soak run on the same spec and flip between them instantly.
Switching Between Types
At the top of the Dashboard panel there's a five-button row — Load / Stress / Spike / Soak / Custom. Clicking any button activates that test type and swaps the configuration form below to the config stored for that type. The rest of the spec (steps, lifecycle scripts, thresholds) is shared and stays put.
Important: the active test type is what the Start Test button runs. Configure all five if you like, but only the currently selected one is executed when you press Start.
Preset Defaults
When you create a new spec, each test type starts from these defaults. You can change any field afterwards — the defaults are just a sensible starting point per preset.
| Type | VUs | Ramp Up | Ramp Down | Duration | Think Time |
|---|---|---|---|---|---|
| Load | 50 | 30 s | 30 s | 60 s | 1000–5000 ms |
| Stress | 200 | 10 s | 10 s | 60 s | 500–2000 ms |
| Spike | 300 | 2 s | 2 s | 30 s | 500–1000 ms |
| Soak | 30 | 30 s | 30 s | 300 s | 2000–5000 ms |
| Custom | 50 | 30 s | 30 s | 60 s | 1000–5000 ms |
Custom inherits the same base defaults but adds an empty stage list you fill in yourself.
When to Use Each Preset
Load
The default sanity check. Ramp up to a realistic target VU count, sustain for the configured duration, and ramp down. Use this to verify the API handles expected traffic without errors or latency regressions.
Stress
Push beyond the normal envelope to find the breaking point. Higher VU count, shorter ramp, same sustain-and-measure pattern. Run increasingly aggressive Stress tests to discover the point where error rate or P95 crosses your SLO.
Spike
Sudden surge — the default ramps 300 VUs over just two seconds. Useful for testing autoscaling behavior, cold-start latency, and how the system recovers when traffic returns to normal.
Soak
Moderate load sustained for a long time. The default hold is five minutes but you can push it up to an hour. Great for surfacing memory leaks, slow resource exhaustion, or gradual latency drift that only appears after extended runtime.
Custom
For profiles the four presets can't express. You define an explicit list of stages — each stage has a target VU count and a duration in seconds. Stages run in order, and the runner interpolates VUs linearly from the previous stage's target to the current one. Use Custom to model traffic patterns like step loads, sawtooth profiles, or multi-peak scenarios.
Custom Stages Editor
When the active test type is Custom, a Custom Stages section appears below the configuration form. Press Add Stage to append a new stage; each new stage inherits the target VU count of the previous one (or 10 if it's the first), with a 30-second default duration.
Per-stage limits:
- • Target VUs: 0 – 5000
- • Duration: 1 – 3600 seconds
Each row has a trash icon to remove it. There's no upper limit on the number of stages — build a 20-step sawtooth if that's what you need.
Tip: a stage with target VUs of 0 is valid — it ramps the load down to zero, effectively idling the runner before the next stage picks it back up.
Per-Type Configuration Is Never Lost
Every spec stores five complete configurations — one per test type — in its configs map. Switching types in the UI only changes which config is being edited. The others stay untouched on disk until you edit them yourself.
This is intentional: you can design a single spec with a gentle Load profile, an aggressive Stress profile, and a multi-hour Soak profile, then choose which one to run on demand.