Admin runtime

Admin Flash Message Runtime Guide

Admin Flash Message Runtime Guide for moving ZartsAlgo admin, portal, connector, traffic, and reporting data toward prepared MySQL runtime at larger client volume.

Runtime contract

What this part of the migration controls.

Use this page as a practical build note before connecting real customer records, provider data, or high-volume event streams.

Boundary

Keep the runtime boundary narrow

This guide moves one admin, portal, import, or reporting behavior at a time so the public site stays stable.

Owner

Assign a runtime owner

Data engineering owns setup, verification, rollback notes, and the first client-facing explanation.

Risk

Name the failure mode

The main risk is mixing raw provider data, client-visible summaries, and high-volume event records into one unbounded runtime path.

First action

Start with the smallest proof

Create one small read or import path, verify row counts, confirm no public page depends on it, and document rollback before expanding.

Target tables

Keep writes organized.

  • crm_lead_status_history
  • crm_leads
  • data_import_batches
  • data_import_errors
  • data_import_queue
  • data_import_rows
  • data_raw_payload_refs

Verification checklist

Do this before switching runtime reads.

  • Confirm the target database user has only the privileges needed for the current phase.
  • Use prepared statements and named parameters for every runtime query.
  • Keep raw provider payloads, webhook bodies, and high-volume events out of the public page layer.
  • Add a rollback path before switching any admin or portal screen away from JSON demo data.
  • Write the client-facing summary from clean report, metric, and source records instead of exposing internal notes.
  • Check cPanel/phpMyAdmin import results and record the applied SQL file date.
  • Verify mobile rendering after the runtime content appears on a dashboard or portal screen.
  • Keep traffic rollups separate from account, client, project, and report records.

Prepared query plan

Small queries beat one giant dashboard query.

Each screen should load a bounded set of records, with source labels and date ranges included wherever the result becomes client-facing proof.

Query 1

portal client by login

Prepare a short SELECT or INSERT path for this behavior, bind parameters, limit result size, and log the outcome without storing secrets.

Query 2

portal report rollup

Prepare a short SELECT or INSERT path for this behavior, bind parameters, limit result size, and log the outcome without storing secrets.

Query 3

portal source coverage

Prepare a short SELECT or INSERT path for this behavior, bind parameters, limit result size, and log the outcome without storing secrets.