Monitoring
Product Monitoring
Continuous tracking of your product's health. Connect data sources — AI will monitor metrics, detect issues and recommend actions.
Getting Started
- Switch to "Monitoring" mode in the top menu
- Click "+ Add" in the Connections section
- Select a data source and enter API keys
- Data will start syncing every 6 hours
Connections
28 data sources in two groups:
| Group | Sources |
|---|---|
| Global | PostHog, Stripe, GA4, Mixpanel, Amplitude, Sentry, Intercom, Zendesk, Typeform, HubSpot, Salesforce |
| Russia | Yandex Metrica, YooKassa, CloudPayments, Tinkoff, AppMetrica, Hawk, Usedesk, Carrot Quest, WebAsk, Testograf, MyTracker, Robokassa, Yandex Forms, AmoCRM, Bitrix24, Yandex DataLens, SberCRM |
Connection statuses
- ● green = active, data syncing
- ● yellow = setup, awaiting API keys
- ○ gray = not connected
6 Health Pulses
The dashboard shows 6 metric cards. Each card: current value + 7-day trend + 30-day sparkline + status (OK/Warning/Critical).
| Pulse | What it shows | Sources |
|---|---|---|
| Traffic | DAU, WAU, MAU | GA4, Yandex Metrica |
| Behavior | Retention, Churn | PostHog, Mixpanel |
| Revenue | MRR, revenue | Stripe, YooKassa |
| Quality | Errors, latency | Sentry, Hawk |
| Support | Tickets, CSAT | Usedesk, Intercom |
| Feedback | NPS, reviews | WebAsk, Typeform |
Alerts
Set up rules — the system will notify you when metrics deviate:
- Conditions: above, below, % increase, % decrease
- Severity is automatic: Critical (>50% deviation), Warning (>20%), Info
- Deduplication: one alert per 24 hours per rule
- Click "Read" to acknowledge
Example rules
Churn > 5% → WARNING alertError rate increase >20% in 24h → CRITICAL alert
MRR decline >10% per week → WARNING alert
Weekly Reports
Every Sunday AI automatically:
- Collects all metrics for the week
- Detects anomalies (Z-score on 30-day history)
- Analyzes causes via LLM
- Generates recommendations
Reports are available in the sidebar and preview panel.
Anomalies
Detection via statistical Z-score:
| Level | Threshold | Description |
|---|---|---|
| Critical | >3σ | Serious deviation from normal |
| Warning | >2σ | Notable deviation, needs attention |
AI analyzes correlations between metrics to determine probable causes of anomalies.
AI Recommendations
Prioritized actions:
| Priority | Description | Action |
|---|---|---|
| HIGH | Product degradation | Fix immediately |
| MED | Important improvement | Schedule this week |
| LOW | Optimization | When you have resources |
Click "To chat" — AI will analyze the issue in detail in the dialogue.
Chat Integration
When monitoring is active, AI in chat can see your metrics. Ask:
Example questions
"Why is churn increasing?" — AI will respond with data from connected sources"Analyze the error anomaly" — detailed correlation analysis
"What should I do about retention?" — recommendations with specific numbers
Tip
Connect at least 2-3 sources from different categories (analytics + payments + quality) —
this allows AI to find correlations between metrics and provide more accurate recommendations.