Skip to content

scheduling.md

Latest commit

 

History

History
113 lines (67 loc) · 3.02 KB

scheduling.md

File metadata and controls

113 lines (67 loc) · 3.02 KB

Testizer email funnels · Scheduling guide

This document describes how often and in what order to run the funnel script, as well as the general scheduling strategy.

1. Which Tasks Need to Run

Main working task:

  • app.main

    Responsible for:

    • selecting new candidates by tests;
    • adding contacts to Brevo lists;
    • recording funnel entry in the funnel_entries table;
    • syncing certificate purchases with MODX tables.

Reporting task:

  • app.report_conversions

    Used for periodic conversion reports, run manually as needed.

2. Main Task Execution Frequency

Recommended logic:

  • Frequency: every 10-30 minutes, depending on load and requirements.
  • With a small number of tests, you can start with hourly runs.

Reasons:

  • tests and purchases don't happen every second;
  • too frequent runs create unnecessary database load;
  • too infrequent runs reduce "freshness" of funnels and emails.

3. Sequence of Actions Before Moving to Schedule

Before enabling automatic execution:

  1. Verify that .env is configured.

  2. Run the script several times manually with APP_DRY_RUN=true:

    python -m app.main

  3. Check logs in logs/app.log.

  4. Change APP_DRY_RUN to false when script behavior is satisfactory.

  5. Run manually again and verify that:

    • contacts actually appear in Brevo;

    • the funnel_entries table is populated;

    • there are no errors in logs.

Only after this should automatic scheduling be enabled.

4. Setting Up Schedule on Server

A production server can use its own scheduling mechanism:

  • cron or similar system on hosting;
  • hosting control panel with graphical task scheduler;
  • separate task orchestrator if already in use.

General recommendations for any system:

  • the startup command should be the Python task from the project root;
  • virtual environment should be activated before startup;
  • work logs should be written to a file for diagnostics.

Specific schedule implementation depends on infrastructure and can be configured by the server administrator.

5. Example Schedule Logic

Regardless of the system used, the logic can be:

  1. Run the main funnel task every 15 minutes.

    • If the previous run is still running, don't start a new task copy.
    • It's important to avoid parallel runs that work with the same tables.

  2. Once a day (e.g., at night) generate a conversion report manually and track dynamics:

    python -m app.report_conversions --from-date 2024-01-01

  3. Change run frequency if needed when:

    • load noticeably increases;
    • funnel logic changes;
    • closer to real-time synchronization is required.

6. Schedule Control and Debugging

After enabling automatic scheduling, you should:

  • periodically check logs/app.log;
  • monitor if new entries appear in funnel_entries;
  • compare the number of new tests and the number of new funnel entries.

If something goes wrong:

  • temporarily disable automatic execution;
  • run the task manually and analyze logs;
  • restore scheduling only after eliminating error causes.