Use syslog-ng db-parser to parse any structure/unstructured data. Based on ELSA that can process at the rate of 6M/sec.
Ranked in these QuestionsQuestion Ranking
Pro Scales easily
ODE instances are independent of each other, so they don't have to worry about a peer being added/removed. This allows the cluster to grow without any performance hit on the log aggregation. There is no redundancy built-in, but you can always use the forwarder to duplicate data. There is no sharding configuration or any other penalty that comes up with scaling a cluster. The clustering configuration is also very easy where you just list out peers for one of the node in order for it to run a search query on the whole cluster and merge the results. Scales better than any other open source log management tool out there.
Pro Add new parsers as you like
You can add any parser you want to ODE.
Pro Highly customizable
Pro Easy to use
Con Still in beta
Opallios ODE seems to be still in beta, as such there may be issues or missing features which are not yet implemented.