kerrad/prod
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

WIP: Development changes moved from master

Browse Source 
- Modified Grafana dashboards
- Updated message sender and metrics collector
- Added new rate limiting dashboard
- Removed count_tests.py
This commit is contained in:
Andrey
2025-09-05 01:29:28 +03:00
parent 7d08575512
commit 40968dd075
7 changed files with 1584 additions and 93 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1227
infra/grafana/provisioning/dashboards/grafana-rate-limiting-dashboard.json Normal file
View File

File diff suppressed because it is too large Load Diff

157
infra/grafana/provisioning/dashboards/server-dashboard.json
View File

@@ -84,7 +84,7 @@
"unit": "percent"
}
},
"gridPos": {"h": 8, "w": 6, "x": 12, "y": 0}
"gridPos": {"h": 8, "w": 6, "x": 18, "y": 16}
},
{
"id": 4,
@@ -139,7 +139,7 @@
}
}
},
"gridPos": {"h": 8, "w": 12, "x": 0, "y": 8}
"gridPos": {"h": 8, "w": 12, "x": 0, "y": 16}
},
{
"id": 5,
@@ -159,7 +159,7 @@
"unit": "s"
}
},
"gridPos": {"h": 8, "w": 6, "x": 12, "y": 8}
"gridPos": {"h": 8, "w": 6, "x": 12, "y": 16}
},
{
"id": 6,
@@ -186,7 +186,7 @@
"unit": "percent"
}
},
"gridPos": {"h": 8, "w": 6, "x": 0, "y": 16}
"gridPos": {"h": 8, "w": 6, "x": 12, "y": 16}
},
{
"id": 7,
@@ -213,7 +213,154 @@
"unit": "percent"
}
},
"gridPos": {"h": 8, "w": 6, "x": 6, "y": 16}
"gridPos": {"h": 8, "w": 6, "x": 18, "y": 16}
},
{
"id": 8,
"title": "CPU Usage Gauge",
"type": "gauge",
"targets": [
{
"expr": "cpu_usage_percent",
"legendFormat": "CPU %"
}
],
"fieldConfig": {
"defaults": {
"color": {
"mode": "thresholds"
},
"thresholds": {
"steps": [
{"color": "green", "value": null},
{"color": "yellow", "value": 70},
{"color": "red", "value": 90}
]
},
"unit": "percent",
"min": 0,
"max": 100
}
},
"options": {
"orientation": "auto",
"reduceOptions": {
"values": false,
"calcs": ["lastNotNull"],
"fields": ""
},
"showThresholdLabels": false,
"showThresholdMarkers": true
},
"gridPos": {"h": 8, "w": 6, "x": 12, "y": 0}
},
{
"id": 9,
"title": "RAM Usage Gauge",
"type": "gauge",
"targets": [
{
"expr": "ram_usage_percent",
"legendFormat": "RAM %"
}
],
"fieldConfig": {
"defaults": {
"color": {
"mode": "thresholds"
},
"thresholds": {
"steps": [
{"color": "green", "value": null},
{"color": "yellow", "value": 70},
{"color": "red", "value": 90}
]
},
"unit": "percent",
"min": 0,
"max": 100
}
},
"options": {
"orientation": "auto",
"reduceOptions": {
"values": false,
"calcs": ["lastNotNull"],
"fields": ""
},
"showThresholdLabels": false,
"showThresholdMarkers": true
},
"gridPos": {"h": 8, "w": 6, "x": 18, "y": 0}
},
{
"id": 10,
"title": "System Resources Overview",
"type": "timeseries",
"targets": [
{
"expr": "cpu_usage_percent",
"legendFormat": "CPU %"
},
{
"expr": "ram_usage_percent",
"legendFormat": "RAM %"
},
{
"expr": "disk_usage_percent",
"legendFormat": "Disk %"
}
],
"fieldConfig": {
"defaults": {
"color": {
"mode": "palette-classic"
},
"custom": {
"axisLabel": "Usage %",
"axisPlacement": "left",
"barAlignment": 0,
"drawStyle": "line",
"fillOpacity": 20,
"gradientMode": "none",
"hideFrom": {
"legend": false,
"tooltip": false,
"vis": false
},
"lineInterpolation": "linear",
"lineWidth": 2,
"pointSize": 5,
"scaleDistribution": {
"type": "linear"
},
"showPoints": "never",
"spanNulls": false,
"stacking": {
"group": "A",
"mode": "none"
},
"thresholdsStyle": {
"mode": "off"
}
},
"unit": "percent",
"min": 0,
"max": 100
}
},
"options": {
"legend": {
"calcs": [],
"displayMode": "list",
"placement": "bottom"
},
"tooltip": {
"mode": "single",
"sort": "none"
}
},
"gridPos": {"h": 8, "w": 12, "x": 0, "y": 8}
}
],
"time": {

184
infra/grafana/provisioning/dashboards/telegram-bot-dashboards.json
View File

@@ -899,6 +899,190 @@
],
"title": "Database Query Time (P95)",
"type": "timeseries"
},
{
"datasource": {
"type": "prometheus",
"uid": "PBFA97CFB590B2093"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "palette-classic"
},
"custom": {
"axisLabel": "",
"axisPlacement": "auto",
"barAlignment": 0,
"drawStyle": "line",
"fillOpacity": 10,
"gradientMode": "none",
"hideFrom": {
"legend": false,
"tooltip": false,
"vis": false
},
"lineInterpolation": "linear",
"lineWidth": 1,
"pointSize": 5,
"scaleDistribution": {
"type": "linear"
},
"showPoints": "never",
"spanNulls": false,
"stacking": {
"group": "A",
"mode": "none"
},
"thresholdsStyle": {
"mode": "off"
}
},
"mappings": [],
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "green",
"value": null
},
{
"color": "red",
"value": 80
}
]
},
"unit": "short"
},
"overrides": []
},
"gridPos": {
"h": 8,
"w": 12,
"x": 0,
"y": 40
},
"id": 11,
"options": {
"legend": {
"calcs": [],
"displayMode": "list",
"placement": "bottom"
},
"tooltip": {
"mode": "single",
"sort": "none"
}
},
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "PBFA97CFB590B2093"
},
"expr": "topk(5, sum by(content_type) (rate(media_processing_total[5m])))",
"refId": "A"
}
],
"title": "Top Media Types",
"type": "timeseries"
},
{
"datasource": {
"type": "prometheus",
"uid": "PBFA97CFB590B2093"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "palette-classic"
},
"custom": {
"axisLabel": "",
"axisPlacement": "auto",
"barAlignment": 0,
"drawStyle": "line",
"fillOpacity": 10,
"gradientMode": "none",
"hideFrom": {
"legend": false,
"tooltip": false,
"vis": false
},
"lineInterpolation": "linear",
"lineWidth": 1,
"pointSize": 5,
"scaleDistribution": {
"type": "linear"
},
"showPoints": "never",
"spanNulls": false,
"stacking": {
"group": "A",
"mode": "none"
},
"thresholdsStyle": {
"mode": "off"
}
},
"mappings": [],
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "green",
"value": null
},
{
"color": "red",
"value": 80
}
]
},
"unit": "bytes"
},
"overrides": []
},
"gridPos": {
"h": 8,
"w": 12,
"x": 12,
"y": 40
},
"id": 12,
"options": {
"legend": {
"calcs": [],
"displayMode": "list",
"placement": "bottom"
},
"tooltip": {
"mode": "single",
"sort": "none"
}
},
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "PBFA97CFB590B2093"
},
"expr": "histogram_quantile(0.5, rate(file_download_size_bytes_bucket[5m])) by (content_type)",
"refId": "A",
"legendFormat": "{{content_type}} P50"
},
{
"datasource": {
"type": "prometheus",
"uid": "PBFA97CFB590B2093"
},
"expr": "histogram_quantile(0.95, rate(file_download_size_bytes_bucket[5m])) by (content_type)",
"refId": "B",
"legendFormat": "{{content_type}} P95"
}
],
"title": "File Download Size Distribution",
"type": "timeseries"
}
],
"refresh": "5s",

18
infra/monitoring/message_sender.py
View File

@@ -74,7 +74,6 @@ class MessageSender:
if self.last_status_time is None:
logger.info(f"should_send_status: last_status_time is None, отправляем статус")
self.last_status_time = now
return True
# Вычисляем разницу в минутах
@@ -84,7 +83,6 @@ class MessageSender:
# Проверяем, что прошло N минут с последней отправки
if time_diff_minutes >= self.status_update_interval_minutes:
logger.info(f"should_send_status: отправляем статус (прошло {time_diff_minutes:.1f} минут)")
self.last_status_time = now
return True
logger.info(f"should_send_status: статус не отправляем (прошло {time_diff_minutes:.1f} минут)")
@@ -92,7 +90,12 @@ class MessageSender:
def should_send_startup_status(self) -> bool:
"""Проверка, нужно ли отправить статус при запуске"""
return self.last_status_time is None
# Отправляем статус при запуске только если он еще не был отправлен
if self.last_status_time is None:
logger.info("should_send_startup_status: отправляем статус при запуске")
return True
logger.info("should_send_startup_status: статус уже был отправлен, пропускаем")
return False
def _get_disk_space_emoji(self, disk_percent: float) -> str:
"""Получение эмодзи для дискового пространства"""
@@ -257,7 +260,14 @@ Read: <b>{system_info['disk_read_speed']}</b> | Write: <b>{system_info['disk_wri
return False
status_message = self.get_status_message(system_info)
return await self.send_telegram_message(self.group_for_logs, status_message)
success = await self.send_telegram_message(self.group_for_logs, status_message)
# Обновляем время последней отправки только при успешной отправке
if success:
self.last_status_time = datetime.now()
logger.info("send_status_message: время последней отправки обновлено")
return success
except Exception as e:
logger.error(f"Ошибка при отправке статуса: {e}")