Use only parquet files for events

This commit is contained in:
Dan Milne
2025-12-03 17:16:38 +11:00
parent 032243ba6a
commit 693851f664
12 changed files with 673 additions and 165 deletions

View File

@@ -0,0 +1,25 @@
# frozen_string_literal: true
# Background job to consolidate completed hour into day file
# Runs at :05 past each hour (e.g., 01:05, 02:05, etc.)
# Merges the previous hour's data into the day file and deletes the hour file
class ConsolidateParquetHourlyJob < ApplicationJob
queue_as :default
def perform
service = AnalyticsDuckdbService.instance
# Consolidate the previous hour (not current hour, which is still being written)
previous_hour = 1.hour.ago
Rails.logger.info "[Parquet Consolidate] Starting hourly consolidation for #{previous_hour.strftime('%Y-%m-%d %H:00')}"
service.consolidate_hour_to_day(previous_hour)
Rails.logger.info "[Parquet Consolidate] Hourly consolidation complete"
rescue StandardError => e
Rails.logger.error "[Parquet Consolidate] Hourly job failed: #{e.message}"
Rails.logger.error e.backtrace.join("\n")
raise # Re-raise to mark job as failed in Solid Queue
end
end