feat: NCS-73 Refine Gatlin tests to reflect ordinary user behaviour #7
Reference in New Issue
Block a user
Delete Branch "feat/NCS-73"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Added realistic user behaviour
Smoke Test
1 user, runs once.
Executes the full user journey exactly once. The goal is a fast sanity check — if this passes, the system is up
and the critical path (register, login, import, move, resign) works end-to-end. Run this before any heavier
test.
Load Test
Ramps up to maxUsers (default 10) over rampDuration (default 60s).
Simulates normal growing traffic. Users are added gradually, each completing the full journey. This reveals how
the system behaves under typical production load and establishes a performance baseline (response times, error
rates) to compare other tests against.
Stress Test
Starts at startUsers (default 2), adds usersIncrement (default 2) users every stepDuration (default 30s) for
steps (default 2) steps, with ramps between.
Deliberately pushes the system beyond normal capacity in controlled steps. Each step holds load steady before
increasing, so you can pinpoint exactly at which user count the system starts degrading or failing — the
breaking point.
Endurance Test
Holds concurrentUsers (default 3) constant for duration (default 300s).
Keeps a steady number of users playing full games repeatedly over a long period. This is the most important test
for detecting slow degradation — memory leaks, connection pool exhaustion, database lock buildup — things that
only appear after sustained use rather than under peak load.
Spike Test
Baseline → sudden burst of spikeUsers (default 15) → 5s pause → back to baseline.
Simulates an unexpected traffic surge (e.g. a news article or viral moment). Tests whether the system can absorb
a sudden large wave of users registering and starting games simultaneously, and whether it recovers cleanly
once the spike subsides.