Visualize Trace analytic

[tzach]

Trace.py use the trace API to generate useful profiling information.
https://github.com/cloudius-systems/osv/wiki/Trace-analysis-using-trace.py

A visualization of the same information should be provided.
The following example data, should be presented as a horizontal bar chart, where each level can be dynamically extend/collapse.

$ trace prof
(100.00%, #31742) All
 |-- (79.68%, #25291) zfs_kmem_alloc
 |    |-- (16.95%, #5380) zio_create
 |    |    |-- (8.96%, #2844) dmu_buf_hold_array_by_dnode
 |    |    |    |-- (8.86%, #2811) zfs_get_data
 |    |    |    |    zil_lwb_commit
 |    |    |    |    zil_commit
 |    |    |    |    zfsvfs_teardown
 |    |    |    |    sys_umount2
 |    |    |    |   
 |    |    |    |-- (0.10%, #32) dmu_buf_hold_array
 |    |    |    |    space_map_sync
 |    |    |    |    metaslab_sync
 |    |    |    |    vdev_sync
 |    |    |    |    spa_sync
 |    |    |    |   
 |    |    |    \-- (0.00%, #1) space_map_load
 |    |    |         metaslab_activate
 |    |    |         metaslab_group_alloc
 |    |    |         zio_dva_allocate
 |    |    |         zio_execute
(...)

[tzach]

Example for a similar visualization from the Java domain, including time and percent of each function, as well as interactive tree widget

[dzautner]

I tried to connect to OSv through VisualVM remotely to help myself a bit during the development but was unable to. Any chance you have some tips about it?

[penberg]

@dzautner https://github.com/cloudius-systems/osv/wiki/Debugging-OSv#profiling-java-applications

If that doesn't help, ping @tgrabiec on the mailing list 😉

[dzautner]

Thanks! forgot to add com.sun.management.jmxremote.authenticate=false, it's working now (:

[dzautner]

I think we should close this issue or at least remove the high priority tag considering the latest discussion about the feature?

[tzach]

@dzautner we still want this feature, we simply do not know how to do it yet

[dzautner]

@tzach my mistake then, sorry