IranArze — Automated Job Deadline Notification
End-to-end Telegram automation that extracts active opportunities, calculates Jalali deadlines, publishes a consolidated daily digest, and sends urgent reminders without routine manual content operations.
Context
2026 · Stopped after real production operation · Maturity: Production.
Problem
Registration deadlines were scattered and easy to miss, while manually monitoring and republishing opportunities was repetitive and fragile.
Users
A Telegram channel that grew to approximately 500 members during the recorded operating period.
My Role
Independent product builder and operator; end-to-end ownership.
Responsibilities
- Defined the daily-digest versus urgent-alert model, urgency levels, message structure, membership gate, and no-manual-content operating model.
- Designed duplicate prevention and operational visibility as product requirements rather than implementation details.
Constraints
- Source HTML structure could change and break parsing.
- Secrets in archived snapshots must be rotated and moved to environment variables before any public code release.
- The service stopped because of infrastructure/communication disruption, not because of product non-use.
Product Decisions
- Separate predictable daily reporting from urgent deadline alerts.
- Make publishing idempotent to avoid duplicates.
- Expose system health so unattended automation can be operated responsibly.
Delivery & Technical Approach
- Built the PHP backend, scraper, Telegram integration, Cron jobs, webhooks, Jalali date logic, idempotency, and JSON persistence.
- Built the operational dashboard, structured logs, health endpoint, cPanel deployment, and production maintenance flow.
Product flow
- Web source
- Scrape & parse
- Calculate Jalali deadline
- Classify urgency
- Duplicate check
- Scheduled Telegram delivery
- Logs / dashboard / health monitoring
Evidence / Outcome
- Approximately 500 channel members.
- Approximately two months of real automated operation.
- Daily publishing and urgent reminders ran without routine manual content operations.
What I Learned
- For automation products, observability and idempotency are user-facing reliability features.
- A useful notification model needs explicit urgency semantics, not just more alerts.
Evidence Boundary
Production refers to the recorded two-month operating period. The service is currently stopped; secrets and credentials are not published.
Next Iteration
Make parsing-health checks source-aware, remove duplicate webhook variants, and deploy with rotated secrets and environment-based configuration.