Hi, I am prospective student for GSoC, currently on the first year of my master's degree and interested in the performance of distributed systems. I'm interested in the End-to-End visualization project and have been talking to the mentor of the project Ali Maredia about it. The project would build on previous efforts that use the blkin library [1] and Twitter's Zipkin [2] to trace and visualize the performance of the components that are used to serve Ceph requests. The traces form trees of spans annotated with timestamps for tracepoints within each request, which are useful to analyze timing breakdowns of correlated calls between different components. A plan has been run through with Ali to rebase blkin tracing to the master branch and fix bugs for the tracing and for the plugin that translates the traces to Zipkin's format. The result would be documented and integrated with the Ceph Benchmarking Tool (CBT), to generate tracepoints while the benchmarks are running and to show end-to-end latencies with Zipkin. It can also be integrated into Teuthology, to generate tracepoints if defined in a job. Any suggestions from the community are welcome! In particular, ideas about the parts of Ceph that would be of interest to be further instrumented for tracing, as well as ideas about the way such end-to-end tracing functionality could help Ceph's development processes. Best regards, Victor Araujo [1] https://github.com/marioskogias/blkin [2] https://github.com/openzipkin/zipkin -- To unsubscribe from this list: send the line "unsubscribe ceph-devel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
