You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Computer1 with rsyslog
Computer2 with rsyslog and logstash
Computer1 is sending data with rsyslog via rsyslog-tls to Computer2, Computer2 receive the encrypted data with rsyslog, decrypt the data and send it to logstash via the loopback.
Computer1 is overloaded with logger during 30 minutes (logs are stored in the RAM).
Step2:
Then the service(s) is/are started on Computer2: for test1 logstash; for test2 logstash and rsyslog.
Result:
Test1:
Test2:
As we can see in Test1 we receive 2000 events/sec, but in Test2 we receive 20000 event/s.
We can't say it's because decoding TLS is taking a lot of resources, because in the second test we are using rsyslog on the same machine as logstash (sort of ssl offloading via rsyslog).
The text was updated successfully, but these errors were encountered:
Hello, it seems that the SSL/TLS option is slowing down the input packet rate, I've done two tests:
1/ Test 1
In this test there are two computers:
Computer1 with rsyslog
Computer2 with logstash (In this test logstash is running with
ssl_enable => "true"
)Computer1 is sending data with rsyslog via rsyslog-tls to Computer2 with logstash.
Schema:
computer1 (rSyslog) ---TLS---> Computer2 (Logstash)
2/ Test 2
In this test there are two computers:
Computer1 with rsyslog
Computer2 with rsyslog and logstash
Computer1 is sending data with rsyslog via rsyslog-tls to Computer2, Computer2 receive the encrypted data with rsyslog, decrypt the data and send it to logstash via the loopback.
Schema:
computer1 (rSyslog) ---TLS---> Computer2 (rSyslog ---No TLS---> Logstash)
Condition of the test:
Step1:
Computer1 is overloaded with logger during 30 minutes (logs are stored in the RAM).
Step2:
Then the service(s) is/are started on Computer2: for test1 logstash; for test2 logstash and rsyslog.
Result:
Test1:
Test2:
As we can see in Test1 we receive 2000 events/sec, but in Test2 we receive 20000 event/s.
We can't say it's because decoding TLS is taking a lot of resources, because in the second test we are using rsyslog on the same machine as logstash (sort of ssl offloading via rsyslog).
The text was updated successfully, but these errors were encountered: