Firebase Analytics Error Tracking Dashboard

Data flows from BigQuery through Cloud Functions into Firestore, with a Tailwind/Chart.js frontend pulling directly from Firestore via the Firebase JS SDK.

The dashboard code is available on GitHub.

The Problem

Firebase Analytics logs SDK errors — invalid event names, oversized parameter values, reserved property names — as firebase_error events. The raw data is accessible in BigQuery, but there is no built-in view for spotting trends across multiple properties simultaneously or detecting when a particular error code starts spiking. The only way to catch a regression early is to query BigQuery manually, which this aims to solve.

How It Works

Data pipeline

A BigQuery SQL query scans each property’s events_* table for firebase_error events and groups the results by date and error code to produce daily counts. The date range is passed in as a variable, so the same query works for both the daily sync and longer backfills. Two Cloud Functions run it:

• A scheduled function runs at 4am daily, syncing a rolling 3-day window (to capture late-arriving Firebase export data) and writing results to Firestore using batch merge: true writes keyed by date_propertyName_errorCode.
• A callable function is triggered from the dashboard and runs a configurable historical backfill — up to 30 days — useful for initial setup or recovering from sync gaps.

Frontend

The dashboard provides four main features:

• Date range filter with a Firestore date-range query on apply.
• Error trend chart — a Chart.js line chart per property, with click-to-isolate so any error code tag in the table updates the chart to show just that property/code combination.
• Breakdown table — togglable between a By Property hierarchy (property → error code → date) and a By Error Code hierarchy (error code → property → date), with expand/collapse tree rows at each level and totals throughout.
• Error code glossary — an inline reference panel covering all Firebase Analytics SDK error codes with plain-English descriptions, so there is no need to leave the dashboard to interpret a code.

Tech Stack

• Cloud Functions v2 (Node.js, scheduled + callable)
• BigQuery (Firebase Analytics export, parameterised queries)
• Firestore (flat document store, batch writes)
• Firebase Hosting
• Firebase Auth (access control)
• Tailwind CSS
• Chart.js

The dashboard code is available on GitHub.