Hi all, I'm investigating efficient logging mechanisms to use in a highly parallel and distributed application. The application currently uses plain text logging. We need logging enabled at all times, although potentially at different log levels. Having to manually manage a tracer is not really an option.
I think using CTF makes sense, since performance and disk space are concerns. One approach would be to use LTTng and having the application interface with its C API to enable tracing on startup and manage the trace. Are there better ways to do this? I was wondering whether logging using libbabeltrace would be a good fit for this use case. I also thought of barectf, but the current limitations [1] make me think it might not be a good fit. Any thoughts or advice on this? Thanks, Mohamad [1] From the Github page: "tracing functions need to be traced from the same thread" _______________________________________________ lttng-dev mailing list lttng-dev@lists.lttng.org https://lists.lttng.org/cgi-bin/mailman/listinfo/lttng-dev