August 2018 Summaries

Firebase announced the general availability of its open-source Admin SDK for .NET, the fifth Firebase Admin SDK variant, enabling server-side applications to access Firebase services alongside the existing web and mobile client SDKs. Supporting .NET Framework 4.5 or later and .NET Standard 1.5 or later, the SDK can be initialized with Firebase service account credentials or automatically discovered Google Cloud application default credentials when running on Google Cloud Platform. Its initial functionality focuses on Firebase Authentication, allowing developers to create custom tokens that integrate external user stores and authentication systems and to verify Firebase ID tokens for secure backend authorization. Firebase invites developers to contribute through GitHub by reporting issues, requesting features, and submitting pull requests.

Firebase Remote Config introduced change history to help teams track, inspect, and reverse configuration updates without manually maintaining records. The feature retains up to 300 versions of a project’s Remote Config values for 90 days, showing changes to parameters and conditions, the person and time associated with each update, and whether it was made through the Firebase console or REST API. Users can compare prior versions and roll back with a button, which creates a new version based on the selected earlier configuration rather than overwriting history. Change history is available through both the Firebase Console and Remote Config REST API, with examples for retrieving recent versions in Node.js, Java, and Python, and is intended to improve transparency and confidence for teams managing live app settings.

Firebase Crashlytics now supports exporting raw crash data to BigQuery, allowing developers to perform deeper analysis beyond the dashboard’s 90-day data-retention period. Exports can be enabled daily for individual apps or projects and include stack traces, logs, custom keys, and other crash details, while giving teams control over retention, deletion, and exports in CSV, JSON, or Avro formats for needs such as GDPR data requests. BigQuery enables developers to investigate crash patterns using custom metadata and Analytics breadcrumbs, such as identifying unstable experiment variants or game levels with high crash rates, and results can be visualized through Data Studio or other business intelligence tools. Existing Fabric users can access the feature by linking their apps to Firebase.

Firebase Hosting introduced support for managing multiple hosting sites within a single Firebase project, allowing related sites such as an admin portal and blog to share resources like Firestore or Realtime Database while being administered through one console and deployed through the CLI. Available on the Blaze plan, the feature lets developers configure several hosting targets in a firebase.json array, associate each target with a specific site through the firebase target:apply hosting command, and deploy or serve individual targets as needed; separate projects are still recommended for staging and production environments with distinct Firebase resources. The update also adds Delta Uploads, which use a local .firebase folder to identify and process only new, changed, or deleted files during deployment, improving performance especially for large sites with many unchanged assets. Both capabilities require Firebase CLI version 4.2.0 or later, and custom domains can be used for subdomains because firebaseapp.com subdomains are not supported.

Firebase In-App Messaging is a tool for guiding users who are actively using a mobile app toward meaningful actions such as subscribing, making purchases, viewing content, or completing game levels. Complementing push notifications delivered through Firebase Cloud Messaging, it supports customizable banner, modal, and image-based messages that can match an app’s branding and user journeys without requiring a new app release. Messages can be triggered through Google Analytics for Firebase events and targeted using profile details, current behavior, and Firebase Predictions about likely future behavior, such as churn or spending likelihood. The service tracks campaign impressions, clicks, and conversions, enabling developers to assess and refine results. Suggested applications include rewarding milestones, incentivizing conversions, promoting new features or content, encouraging social sharing, and offering contextual tips.

Firebase announced platform updates aimed at helping mobile teams grow user engagement, improve app quality, and build scalable backends on Google Cloud. New Firebase In-App Messaging enables targeted, customizable messages for active users using Analytics and Predictions data, while an updated Firebase Cloud Messaging dashboard centralizes notification metrics across platforms and Remote Config adds version history and rollback support. Crashlytics gains BigQuery export for deeper crash analysis and a Jira Software integration for issue tracking. Cloud Firestore increases capacity to 10,000 writes per second and 1 million concurrent users, Cloud Functions reaches general availability with an SLA, and both services expand to additional European and Asian regions. Firebase Hosting will support multiple websites per project and faster deployments by uploading only changed files, while the redesigned console Project Overview combines analytics, crash, performance, messaging, testing, and service-health information into a unified dashboard.

Cloud Functions for Firebase reached general availability at Cloud Next 2018 alongside new deployment configuration options in the Firebase CLI. Developers can now use the beta Node.js 8 runtime, which provides V8 performance improvements and native ECMAScript 2017 async/await support, by updating the Firebase CLI and firebase-functions package and specifying Node 8 in package.json; TypeScript projects can also target ES2017 to avoid transpiling async/await into generators. The update additionally allows per-function region selection, memory allocation, and timeout settings through the firebase-functions API, helping teams place workloads closer to users and tailor resources to specific tasks. Although Google Cloud Functions added Python support, that runtime was not yet supported through the Firebase CLI.

Cloud Firestore introduced improved array support to address the limitations and concurrency risks of modifying arrays by index in multi-user applications. Developers can now query documents for values contained in arrays using `array-contains`, avoiding earlier map-based workarounds and supporting non-string array values. New `arrayUnion` and `arrayRemove` operations provide set-like updates by adding only values not already present and removing all matching values rather than targeting specific positions. The updates also expand security rules, allowing access controls based on whether an array contains a specified element, such as permitting queries only when a user appears in a document’s viewers array. These capabilities require the latest Firestore client SDKs.

Cloud Firestore updates announced around Cloud Next 2018 include controls to disable automatic single-field indexing, helping developers avoid unnecessary storage and performance costs for large arrays or maps, prevent the 20,000 indexed-fields-per-document limit, and bypass the 500-writes-per-second restriction associated with sequentially indexed values such as timestamps, though unindexed fields cannot be queried. New database hosting locations, initially Frankfurt and South Carolina, aim to improve latency and support data-residency requirements, with the selected location fixed at project creation. Firestore also adds import and export capabilities through Google Cloud Storage for backups, migration, and copying data across environments, while faster security-rule deployment is intended to reduce activation times from minutes. Backend improvements raise beta limits to 10,000 writes per second per database and 1 million concurrent connections. Finally, Firestore supports Native mode for Firebase and Google Cloud applications and Datastore mode for backward-compatible Cloud Datastore users, adding strong consistency and fewer write and transaction limits while retaining some Datastore feature differences.