You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Testing Regal on a medium large policy repo (116 policy files) and unsurprisingly, it's slow — total time to lint clocked in at about 12 seconds. This is with all the rules enabled, mind you, but since we're likely making that the default, we should try and improve this before a first release. Collecting and reporting metrics will be of great help for further optimizations, but since most of the time is likely always spent evaluating Rego — and each input file is evaluated individually — it seems like an easy win to parallelize linting across all the input files.
An option to turn this off, or to configure the level of parallelism would be nice, but perhaps not needed for the first iteration.
The text was updated successfully, but these errors were encountered:
This only makes the rego evaluation concurrent.
An option to write output to a file has also been added.
Fixes: #113, at least as a first
pass. There is always GOMAXPROCS.
Signed-off-by: Charlie Egan <charlie@styra.com>
Testing Regal on a medium large policy repo (116 policy files) and unsurprisingly, it's slow — total time to lint clocked in at about 12 seconds. This is with all the rules enabled, mind you, but since we're likely making that the default, we should try and improve this before a first release. Collecting and reporting metrics will be of great help for further optimizations, but since most of the time is likely always spent evaluating Rego — and each input file is evaluated individually — it seems like an easy win to parallelize linting across all the input files.
An option to turn this off, or to configure the level of parallelism would be nice, but perhaps not needed for the first iteration.
The text was updated successfully, but these errors were encountered: