Trace Logs - decrease verbosity

Using FoundationDB
1

Hello,

we are preparing to run several foundationdb (Version 7.1+) clusters hopefully in production. we just integrated splunk logging in our cluster design - for the trace logs generated by foundationdb - and it seems that it logs an excessive amount (at least for my preferences)
(i’m getting ~ 6 mil events per hour from a 21 machine cluster/8 processes per machine)

one of the directions that i’ll take is to investigate if there is anything wrong with the cluster. However the question is if there is any way to increase minimum severity of the messages it logs ?

It seems that 80%+ of the messages are severity 10 which it seems it’s some kind of debugging level. I’m sure it might be useful in the future to monitor the state of the cluster but for now i’m just thinking of turning it off.

2

My understanding is that disabling these lower priority events is not recommended because the system actually relies on the presence of some of these events to do certain kinds of communication between processes. This is clearly undesirable and should be fixed, but for now I think that means the level cannot be set above 10.

1 Like