The recordings from Apache Community Over Code EU 2026 are up, and both OpenAI talks cover how they use GreptimeDB for Spark observability: • _*Operating Spark at OpenAI’s Hyperscale Lakehouse*_: per-task metrics mean tens of millions of time series a day, and logs, metrics and events sat in separate systems. They now keep metrics, events and logs in GreptimeDB, read it from Grafana and their Rust Spark History Server, and query the Parquet on S3 with Spark and Trino. From 2945 • _*Beyond the Spark UI*_: a listener in the Spark driver ships job, stage and task records through Grafana Alloy into GreptimeDB, which let them drop one of their most expensive Flink jobs. From 1900 Worth a watch if you run Spark or high-cardinality workloads.