Skip to content


CARET is able to analyze data recorded for several minutes. When analyzing large data with CARET, both memory and time consumed for analysis will be critical issue.

In order not to load unnecessary data, CARET has LTTngEventFilter to exclude selected data from loading data.

LTTngEventFilter has the following methods.

  • init_pass_filter
  • duration_filter
  • strip_filter


With init_pass_filter method, events recorded by runtime tracepoints are not loaded onto memory.



With duration_filter method, you can crop data with targeting time range of recorded data.

LttngEventFilter.duration_filter(duration_s: float, offset_s: float)

duration_filter method has following two arguments.

  • duration_s is duration of target time range
  • offset_s is point of target time range


With strip_filter method, you can crop data with targeting time range of recorded data as well as duration_filter. strip_filter has different arguments from duration_filter.

LttngEventFilter.strip_filter(lsplit_s: Optional[float], rsplit_s: Optional[float])

strip_filter method has following two optional arguments. If you omit either or both of them, default value '0' is given to the optional argument.

  • lstrip_s is start time of cropping range
  • rstrip_s is end point of cropping range

Sample code#

A sample code, where duration_filter is used, is given as below.

from caret_analyze import Lttng, LttngEventFilter

lttng = Lttng('/path/to/ctf', event_filters=[
  LttngEventFilter.duration_filter(10, 5)
]) # Filtering for 10 from 5 seconds

If you want to add another filter, append it to event_filters list.