Conference Schedule - PGConf.EU 2012Back
Using PostgreSQL for storing time-series data
Date: Oct. 26, 2012
As of today, RRDtool is the de-facto standard for storing and graphing time-series data. However, that tool is fairly limited when it comes to large setups requiring redundancy, high-availability, complex analysis of the data or advanced data queries. An RDBMS like PostgreSQL, on the other hand, is very well suited for that kind of purpose. This talk describes and analyses a setup using PostgreSQL as storage backend for storing (high-frequent) performance data and the performance implications on the PostgreSQL setup. We will discuss an approach how to tackle those issues and present benchmark results for those kinds of setups. In the second part of the talk, PostRR (PostgreSQL Round Robin Extension) will be introduced, which provides an extension for PostgreSQL to store data in a round-robin fashion by introducing new data-types for that purpose. Currently, this is in an early stage of development, so there is room for discussion and future directions of the project.