Eventnative Versions Save

What's New

• 🎯 HubSpot destination
• 🚅 Mixpanel batch mode support that improves performance and reduces traffic usage for projects with high events number
• ❄️ Schema Freeze connection option that disable automatic schema management. Read more

Other significant changes

• 🎯 Amplitude destination: added support for Screen events
• 🎯 Intercom, Google Analytics 4, Google Tag manager destinations: improved event payload
• 🔄 Syncs: added option to customize destination table name for every stream
• 🔄 Syncs: respect original source data schema when syncing to data warehouses
• 🪲 Redshift, Postgres destinations: fixed bug when destination table couldn't be updated when database schema was provided in upper case
• 🪲 @jitsu/js #1098 fixed bug: setAnonymousId not working
• 🚀 bulker: general performance improvements for batch uploading
• 🚀 rotor: performance optimization for connection having multiple Functions attached
• 🚀 memory leak fixes for bulker and rotor

[!NOTE] Upgrading from Jitsu 2.4 requires infrastructure changes. Details below.

What's New

• Live Events: Now you can see sync tasks' batch logs right in the Data Warehouse Events tab.
• Rotor: We've tuned the performance for configurations that use multiple functions per connection, making things smoother and faster.

Upgrading from 2.4

We've made a couple of significant changes:

• Live Events: We've moved from Redis to ClickHouse for better scalability and performance.
• Persistent Storage: We're now using MongoDB instead of Redis, which means more robust data handling.

Redis is no longer in the picture.

For Live Events, we didn't set up a migration path because we see it as a nice-to-have, not a must-have.

If Persistent Storage is part of your setup, enable "double reading" by setting up both MONGODB_URL and REDIS_URL. Jitsu will read from both but only write to MongoDB. This way, you won't lose any data during the transition. After a while, when all your data is happily living in MongoDB, you can say goodbye to Redis.

It's best to switch to MongoDB and ClickHouse sooner rather than later. We're planning to drop Redis support after a few more releases, so keep an eye out.

For detailed configuration instructions and further assistance, please se the configuration guide or the reference Compose file.

This hotfix release includes the following changes:

• 🔄 console: fixed bug with Sign In page not working on some setups
• 🔄 console: Live Events now has the toggle to enable Functions debug logs
• ⚙️ GTM destination fix for GTM tag initialization
• ⚙️ rotor: another fix for graceful shutdown

👉 Full changelog

This hotfix release includes the following changes:

• 🔄 console: fixed bug in anonymous telemetry system that could cause error and timeout on some UI operations.
• ⚙️ rotor: token authorization can be enabled for rotor endpoints
• ⚙️ rotor: graceful shutdown now must work properly
• 💅🏾 @jitsu/js. Jitsu processing queue now accepts arrays in addition to callbacks
• unified format of AUTH_TOKENS for all components.

👉 Full changelog

This hotfix release includes the following changes:

• ⚠️ Important: Authorization was not enabled for: bulker, ingest, syncctl services when configured using raw tokens (primarily in docker-compose)
• console: listen for requests on all hostnames (0.0.0.0) - may be necessary for productions setups

This release offers a refinement to self-hosted deployment process among with bunch of other fixes and improvements.

Due to number of changes migration from previous docker-compose setup is not supported

Here is the list of the most significant changes:

✅ Event Ingestion API was moved to the dedicated service written in Go language: jitsucom/ingest ✅ GitHub OAuth app is no longer required to setup Jitsu ✅ Improved performance of jitsucom/rotor service. Now it supports concurrent execution of multiple functions ✅ Simplified setup of Connector Syncs with much shorter Minikube example ✅ jitsucom/console image was significantly reduced in size, from 900M to 172M ✅ Kafka image upgraded to the fresh one bitnami/kafka:3.6.0

This hotfix release includes the following changes:

• Postgres Destination: added support for servers with SNI
• Google Ads source: support for custom WHERE clause for GAQL query
• Salesforce source: added native airbyte version
• Firebase Source: load user's DisplayName and PhotoURL properties
• Airbyte and Singer based sources: correctly detect numeric types in data columns

This hotfix release includes the following changes:

• Fixed Javascript Debugger UI not rendered correctly.
• Facebook Marketing source: API version bumped to 17. Now API version can be overridden with environment variable FACEBOOK_MARKETING_API_VERSION

This hotfix release includes the following changes:

• Snowflake: fix for table cleanup didn't work properly during source syncs
• Sources Sync: support more variants of date formats in source data
• Sources Sync: fix for bug with primary key calculation of objects without explicit key that could lead to data duplication in some cases
• Firebase source: optimize performance of sub-collection syncs
• Airbyte Sources: don't offer dev versions of images that could lead to failed syncs
• Added Google Analytics (GA4) airbyte source
• Added WooCommerce airbyte source.
• Build fixes

This hotfix release includes the following changes:

• Google Ads source: correct data type for videoQuartile-s metrics
• Source Sync Tasks: allow tasks to run for up to 48 hours