Real Time Analytics with PostgreSQL

Many companies are in need of a real-time analytics solution. Events such as website clicks or sensor measurements come in at a high rate and users want real-time insight into this data using a dashboard, which involves a number of scalability problems. A typical approach is to store events data in HDFS, compute different aggregates using Spark, and put the results into Cassandra to let the application access them. In this webinar, we will show that PostgreSQL can perform all of these tasks efficiently by using a combination of cached aggregates and a queue of changes that is periodically flushed into the cache. Using the Github archive data, we will demonstrate that this approach performs well on both the ingestion side and the querying side even for very large datasets.

[@andrewshiner1606]

Very interesting talk. Thank you for sharing.

[@igboman2860]

What about percentile aggregate, that can not be incrementally updated. can it ?