Hi, I've read that the Graylog 0.20 RC1 was released. Now i will begin to test my new Graylog2 infrastructure for stability and performance. Performance is the reason why I'm asking. I googled a bit and couldn't find any benchmark utility, that can generate enough messages to bring my infrastructure to their limit. What i found was a simple bashscript, that generates about 1200 msg/s running on my Intel Core i7 with 8 threads/instances of the same script at 100% CPU load. The sad thing was, my Graylog2 Server running on a Intel Core i5 used about 20% to parse the messages. Of course i can multiply 1200 msg/s by 5, but under heavy load, services will steal each other CPU cycles and in the end effect the server will probably be much slower than originally calculated.
I think for someone, exactly as me, who will prepared to an increasing amount of logs caused by an expanding webserver infrastructure or a DDOS, will need such a utility. What I imagine is an utility, that generates messages with different lengths, different error levels and so on. A additional feature that would help to keep the indexes clean, is that Graylog2 server will automatically generate a temp index for the benchmark. The advantage is that you could simulate real load with sequential disk iops, your existing indexes will not be touched and for statistic nerds, as me, the message counter would not be distorted. Please let me know, if there exist an utility that I've overlooked and if such a utility will be developed by Company behind Graylog2. Cheers :-) -- You received this message because you are subscribed to the Google Groups "graylog2" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/groups/opt_out.
