I don’t want this to become a spam forum, nothing further in my intention than this, but Pascal (firstname.lastname@example.org) just shared some screenshots with us from a customer PoC, which are much more interesting than any internal benchmark we could do in the labs.
These screenshots show the Logger performing in the worst case scenario; when the search query is not matched while inserting events on disk. Two pics are provided with about 2 weeks of separation and the hardware is a L3000 appliance.
- ~3 seconds for ~ 2.7 billion events.
- ~4.5 seconds for ~ 3.2 billion events.
These are the final steps of benchmarking and debugging and this new logger iteration should be available before end of the year (3.1 probably, not sure about the versioning tho). More on the final benchmark in the future, I’m thinking about interviewing some of our key people for the Community/Labs page and having Carlos talk about this would be great.
Huge kudos to everyone involved, specially to Carlos, Juanma and Mihnea :-)