PostGrep: A library to parse PostgreSQL logs

PostGrep is a collection of Haskell libraries to parse and interpret PostgreSQL logs. The goal of PostGrep is to provide a way for users to easily create visualizations of their logs, and also to expose library components that allow users to create their own log parsing system.

Currently, postgrep is composed of 3 packages:

• postgrep-core: This library implements the basic functionality of converting a log into a list of LogEntrys.
• postgrep-conduit: This very small and simple library exposes functions that should be used with the conduit library. This is the recommended way of parsing logs. It is broken off so postgrep-core doesn't have extra dependencies.
• postgrep-gcharts: This package provides a user-facing executable that parses a log and visualizes the contents with Google Charts. It is also intended to act as a library so users can build their own visualizations.

Usage

For each of these libraries, the only things required by the user should be a file or ByteString with their log, and the log_line_prefix format for their logs. First, postgrep parses the log_line_prefix into a regex-based parser. This is then used to parse log lines and group them together into log entries.

TODO

• Handle syslog and syslog2 lines. So far we only handle stderr style logs.