chore(): docs improvements

README.md

@@ -136,7 +136,19 @@ System metrics (gc, requests, etc.) and several business metrics (`lotty_decide_
 
 ### OTEL
 
-You can enable full observability for backend by setting `OTEL_ENABLED=True` in `.env`, also you need to run `compose.prod.yaml` with `observability` profile in order to deploy full observability stack
+You can enable full observability for backend by setting `OTEL_ENABLED=True` in `.env`, also you need to run `compose.prod.yaml` with `observability` profile in order to deploy full observability stack.
+
+![Grafana Logs](./assets/images/logs_grafana.png)
+![Grafana Traces](./assets/images/tracing.png)
+![Grafana Metrics](./assets/images/metrics_grafana.png)
+
+Also there is some other predefined dashboards for services.
+
+Grafana ([localhost/grafana/](http://localhost:80/grafana/)):
+
+login: `admin`
+
+password: `prooooood`
 
 ## Load testing (k6)
 

artifacts/k6/20260224171916/fixture.json

@@ -0,0 +1 @@
+{"run_id": "r_20260224171916", "flag_id": "c4233665-691f-4913-855e-054cb0b634c7", "flag_key": "k6_r_20260224171916_flag", "experiment_id": "3fae9bd2-d1eb-4417-8dbc-3f48bd99df16", "experiment_status": "running", "owner": "experimenter", "approver": "approver", "subject_attributes": {"country": "US"}}

artifacts/k6/20260224171916/run.env

@@ -0,0 +1,16 @@
+RUN_ID=20260224171916
+K6_IMAGE=grafana/k6:0.50.0
+K6_BASE_URL=http://host.docker.internal
+FLAG_KEY=k6_r_20260224171916_flag
+SUBJECT_COUNTRY=US
+START_RPS=20
+RAMP_UP_RPS=100
+HOLD_RPS=100
+PRE_ALLOCATED_VUS=100
+MAX_VUS=600
+RAMP_UP_DURATION=30s
+HOLD_DURATION=2m
+RAMP_DOWN_DURATION=20s
+THRESHOLD_ERROR_RATE=0.01
+THRESHOLD_P95_MS=250
+THRESHOLD_P99_MS=500

artifacts/k6/20260224171916/summary.json

@@ -0,0 +1,176 @@
+{
+    "root_group": {
+        "id": "d41d8cd98f00b204e9800998ecf8427e",
+        "groups": {},
+        "checks": {
+                "status is 200": {
+                    "fails": 2,
+                    "name": "status is 200",
+                    "path": "::status is 200",
+                    "id": "6210a8cd14cd70477eba5c5e4cb3fb5f",
+                    "passes": 14798
+                },
+                "has one decision": {
+                    "name": "has one decision",
+                    "path": "::has one decision",
+                    "id": "7486472dbb4301cd7f9433accea21ab5",
+                    "passes": 14798,
+                    "fails": 2
+                }
+            },
+        "name": "",
+        "path": ""
+    },
+    "metrics": {
+        "vus_max": {
+            "value": 100,
+            "min": 100,
+            "max": 100
+        },
+        "data_sent": {
+            "rate": 22664.865759829918,
+            "count": 3854582
+        },
+        "iterations": {
+            "count": 14800,
+            "rate": 87.02370665495839
+        },
+        "decide_status_200_rate": {
+            "passes": 14798,
+            "fails": 2,
+            "thresholds": {
+                "rate>0.99": false
+            },
+            "value": 0.9998648648648648
+        },
+        "decide_experiment_assigned_rate": {
+            "passes": 14798,
+            "fails": 2,
+            "value": 0.9998648648648648
+        },
+        "http_req_blocked": {
+            "med": 0.003334,
+            "p(90)": 0.011583,
+            "p(95)": 0.018917,
+            "p(99)": 0.19010065999999978,
+            "max": 16.584178,
+            "avg": 0.013590071283783596,
+            "min": 0.001208
+        },
+        "http_req_waiting": {
+            "avg": 39.85390373709466,
+            "min": 11.816621,
+            "med": 16.489178,
+            "p(90)": 77.800971,
+            "p(95)": 166.82410669999987,
+            "p(99)": 419.18147045999984,
+            "max": 1044.069215
+        },
+        "iteration_duration": {
+            "med": 16.804991,
+            "p(90)": 78.58091970000002,
+            "p(95)": 168.52110554999976,
+            "p(99)": 421.9347988599999,
+            "max": 1048.179228,
+            "avg": 40.33107365702705,
+            "min": 11.977788
+        },
+        "vus": {
+            "value": 0,
+            "min": 0,
+            "max": 36
+        },
+        "http_req_duration": {
+            "max": 1047.518767,
+            "avg": 40.00110370472988,
+            "min": 11.857496,
+            "med": 16.56947,
+            "p(90)": 78.08279640000002,
+            "p(95)": 167.87539784999998,
+            "p(99)": 420.48415555999975,
+            "thresholds": {
+                "p(95)<250": false,
+                "p(99)<500": false
+            }
+        },
+        "http_req_connecting": {
+            "med": 0,
+            "p(90)": 0,
+            "p(95)": 0,
+            "p(99)": 0,
+            "max": 1.364963,
+            "avg": 0.003630033851351352,
+            "min": 0
+        },
+        "http_req_duration{expected_response:true}": {
+            "med": 16.5694485,
+            "p(90)": 78.08813720000002,
+            "p(95)": 167.87621554999998,
+            "p(99)": 420.52659067999974,
+            "max": 1047.518767,
+            "avg": 40.00300859345872,
+            "min": 11.857496
+        },
+        "http_req_sending": {
+            "p(90)": 0.040708100000000004,
+            "p(95)": 0.062042,
+            "p(99)": 0.21800307999999996,
+            "max": 35.933239,
+            "avg": 0.03309161581081086,
+            "min": 0.004334,
+            "med": 0.015917
+        },
+        "decide_requests_total": {
+            "count": 14800,
+            "rate": 87.02370665495839
+        },
+        "http_req_tls_handshaking": {
+            "avg": 0,
+            "min": 0,
+            "med": 0,
+            "p(90)": 0,
+            "p(95)": 0,
+            "p(99)": 0,
+            "max": 0
+        },
+        "http_req_receiving": {
+            "p(90)": 0.125209,
+            "p(95)": 0.23725719999999992,
+            "p(99)": 1.3213682599999996,
+            "max": 27.769673,
+            "avg": 0.11410835182432402,
+            "min": 0.006667,
+            "med": 0.04525
+        },
+        "decide_request_duration_ms": {
+            "min": 11.857496,
+            "med": 16.56947,
+            "p(90)": 78.08279640000002,
+            "p(95)": 167.87539784999998,
+            "p(99)": 420.48415555999975,
+            "max": 1047.518767,
+            "avg": 40.00110370472988
+        },
+        "http_req_failed": {
+            "fails": 14798,
+            "passes": 2,
+            "thresholds": {
+                "rate<0.01": false
+            },
+            "value": 0.00013513513513513514
+        },
+        "http_reqs": {
+            "count": 14800,
+            "rate": 87.02370665495839
+        },
+        "checks": {
+            "fails": 4,
+            "passes": 29596,
+            "value": 0.9998648648648648
+        },
+        "data_received": {
+            "count": 8109462,
+            "rate": 47683.475825508925
+        }
+    }
+}

compliance-matrix.md

@@ -53,7 +53,7 @@
 | `3.7` | `B9-2` | Нужен отдельный liveness probe | `src/backend/api/urls.py` (`/health`) | Runtime: `curl -i /health` | Поднятый backend | подтверждено |
 | `D.5(B9)` | `B9-3` | Без метрик нет наблюдаемости hot-path | `src/backend/config/settings/base.py`, `src/backend/apps/decision/services.py`, `src/backend/api/v1/events/endpoints.py` | Runtime: `curl /metrics` | Поднятый backend | подтверждено |
 | `D.5(B9)` | `B9-4` | Неструктурированные логи сложны для алертов и анализа | `src/backend/config/settings/base.py` (json formatter, django-guid) | Запуск в non-debug и просмотр stdout | Конфигурация `DJANGO_DEBUG=false` | подтверждено |
-| `D.5(B9)` | `B9-6` | Рост трафика/данных может деградировать latency даже после оптимизаций | `src/backend/apps/decision/services.py`, `src/backend/apps/reports/services.py`, `src/backend/apps/events/tasks.py`, `ADR/04-decisions.md` (P1-P4) | Код-ревью + live-demo под нагрузкой | Сценарий увеличенного трафика и наблюдение очереди/latency | частично (live-demo) |
+| `D.5(B9)` | `B9-6` | Рост трафика/данных может деградировать latency даже после оптимизаций | `infrastructure/k6/decide.js`, `infrastructure/k6/run-decide.sh`, `infrastructure/k6/README.md`, `src/backend/apps/decision/services.py`, `src/backend/apps/reports/services.py`, `src/backend/apps/events/tasks.py`, `ADR/04-decisions.md` (P1-P4) | Нагрузочный прогон `./infrastructure/k6/run-decide.sh` + анализ `artifacts/k6/<RUN_ID>/summary.json` (p95/p99, error rate, req/s); фактические прогоны: `artifacts/k6/smoke_k6`, `artifacts/k6/20260224171916`; в прогоне `20260224171916`: `14_800` запросов, `87.02 req/s`, `p95=167.88 ms`, `p99=420.48 ms`, `error_rate=0.0135%` (2/14800) | Поднятый compose-стек и артефакты `artifacts/k6/<RUN_ID>/{summary.json,run.env,fixture.json}` | частично (live-demo) |
 | `D.5(B9)` | `B9-7` | Без индексов и оптимизаций горячие запросы дорожают | `src/backend/apps/experiments/models.py`, `src/backend/apps/events/models.py`, `src/backend/apps/guardrails/models.py`, `src/backend/apps/notifications/models.py`, `src/backend/apps/learnings/models.py` | Схема моделей и миграций | БД-схема проекта | подтверждено |
 | `D.5(B10)` | `B10-1` | Отсутствие автоматического линтинга снижает качество | `src/backend/justfile`, `src/backend/pyproject.toml`, `.gitlab-ci.yml` | `cd src/backend && just lint` | Dev dependencies | подтверждено |
 | `D.5(B10)` | `B10-2` | Отсутствие форматирования повышает шум в diff | `src/backend/justfile`, `src/backend/pyproject.toml`, `.gitlab-ci.yml` | `cd src/backend && just format` | Dev dependencies | подтверждено |

infrastructure/k6/run-decide.sh

@@ -9,8 +9,8 @@ K6_IMAGE="${K6_IMAGE:-grafana/k6:0.50.0}"
 K6_BASE_URL="${K6_BASE_URL:-http://host.docker.internal}"
 
 START_RPS="${START_RPS:-20}"
-RAMP_UP_RPS="${RAMP_UP_RPS:-200}"
+RAMP_UP_RPS="${RAMP_UP_RPS:-100}"
-HOLD_RPS="${HOLD_RPS:-200}"
+HOLD_RPS="${HOLD_RPS:-100}"
 PRE_ALLOCATED_VUS="${PRE_ALLOCATED_VUS:-100}"
 MAX_VUS="${MAX_VUS:-600}"
 RAMP_UP_DURATION="${RAMP_UP_DURATION:-30s}"