v1.2.6

This release contains the Parquet support for the ingestion (#62) , allowing Talaria to directly read and ingest Parquet files into its storage layer.

v1.2.5

This release contains the upgrade of ORC reader in order to fix the reader of float64 as well as a fix for some of our unit tests #61.

v1.2.4

This release contains various bug fixes.

• Added support for map[string]interface{} to being auto-serialized into JSON for Golang client #54
• Fix filter to return nil when data is filtered out #53 by
• Fix sub streaming not starting #52

v1.2.3

This release adds real-time sub-streaming capability to Talaria. This can be configured per-table by using streams configuration.

Example configuration

tables:
  eventlog:
    streams:
      - pubsub:
          project: my-gcp-project
          topic: my-topic
          filter: "gcs://my-bucket/my-function.lua"
          encoder: json
...

Changelog

• Streaming capability for Sinks.
• Added Google Pub/Sub as a possible sub-streaming sink.

v1.2.2

This release adds WithLoadBalancer option to the Golang client of talaria, allowing to specify clientside load balancing strategy.

v1.2.1

Changelog

This release fixes the following changes:

• When querying, empty frames will be ignored which previously resulted in a panic.
• When ingesting a varchar column, allow for schema re-mapping so that it can be converted to JSON, int32 or other types.
• Golang client for Talaria now includes kubernetes resolver, allowing to Dial() with kubernetes schema. See: https://github.com/sercand/kuberesolver

v1.2.0

This version adds a support for multiple tables and ultimately introduces ability to re-partition data automatically by multiple different dimensions, during the ingestion process. For example, data can be partitioned by event, booking, and user_id by providing a following configuration:

tables:
  events:
    hashBy: "event"
    sortBy: "ingested_at"
  bookings: 
    hashBy: "booking_id"
    sortBy: "ingested_at"
  users: 
    hashBy: "user_id"
    sortBy: "ingested_at"

Breaking changes:

• Environment variable configuration TALARIA_CONF prefix is replaced with TALARIA
• Table configuration has been revamped (see /config/sample_config.yaml)
• Removed the ability to configure logs and nodes tables as they were never actually changed
• Compaction configuration is now per table

v1.1.18

• Added Talaria Sink in order to re-publish compacted data into a second Talaria

v1.1.17

• Support JSON string input if desired static schema column is JSON #39

v1.1.16

• Increased compression chunk in ORC in order to support large strings.