docs: improve stress test doc

Author: maxceem

test/stress/README.md

@@ -1,6 +1,20 @@
+# Stress test
+
+## Overview
+
+This test is checking if any error would happen if we create, updated and delete milestones in **the same** timeline document in Elasticsearch `timelines` index at the same time. We are sending multiple Kafka events in parallel to trigger create, updated and delete milestone operations.
+
+1. First, the script should create initial data in the ES index:
+   - create one timeline in the `timeline` index with `2*STRESS_BASIC_QTY` milestones (it's important that all milestones belong to the same timeline)
+
+2. After initial data is created start an actual stress test by sending `3*STRESS_BASIC_QTY` Kafka messages in parallel:
+   - `STRESS_BASIC_QTY` Kafka messages to delete half of the initially created milestones
+   - `STRESS_BASIC_QTY` Kafka messages to update another half of initially created milestones
+   - `STRESS_BASIC_QTY` Kafka messages to create `STRESS_BASIC_QTY` new milestones in the same timeline
+
 ## Configuration
 
-* `STRESS_BASIC_QTY`: As requested in challenge description
+* `STRESS_BASIC_QTY`: The basic number of objects to use in stress test.
 
 * `STRESS_TESTER_TIMEOUT`: Number of seconds to wait after queueing create/update/delete requests and before validating data. Default is 80s, which is enough for `STRESS_BASIC_QTY=100`. This might have to be increased if `STRESS_BASIC_QTY` is higher than 100.
 
@@ -21,5 +35,3 @@
   ```
   npm run test:stress
   ```
-
-I would save ids of milestones not updated/created/deleted to `test/stress/stress_test_errors.json`