Rules matching perf improvements #2210
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
mstemm
force-pushed
the
rules-matching-perf-improvements
branch
2 times, most recently
from
80ef37f to
19483c1
Compare
mstemm
force-pushed
the
rules-matching-perf-improvements
branch
from
19483c1 to
879b8d4
Compare
leogr
reviewed
Sep 15, 2022
userspace/engine/falco_engine.cpp
Outdated
| return add_source(source, filter_factory, formatter_factory, ruleset_factory); | ||
| size_t idx = add_source(source, filter_factory, formatter_factory, ruleset_factory); | ||
|
|
||
| if(source == "syscall") |
There was a problem hiding this comment.
Suggested change
| if(source == "syscall") | |
| if(source == falco_common::syscall_source) |
When doing some testing of falco on very high event volumes (> 1.5M events/second), I found that the time taken to look up a falco_source struct had a non-negligible contribution to cpu usage. So instead of looking up the source from the source_idx every time, separately save the source for syscalls in the falco_engine object directly. The separately saved copy is only used once someone calls add_source with source="syscall". Signed-off-by: Mark Stemm <mark.stemm@gmail.com>
Instead of using a falco_rule struct on the stack, use a single value inside the falco_source struct. It's mutable as find_source returns a const struct. At very high event volumes (> 1M syscalls/second), even the tiny time it takes to create/destroy the struct starts to add up, and this switch has some small cpu savings. Signed-off-by: Mark Stemm <mark.stemm@gmail.com>
mstemm
force-pushed
the
rules-matching-perf-improvements
branch
from
879b8d4 to
a00443b
Compare
|
/milestone 0.33.0 |
|
LGTM label has been added. Git tree hash: d7a242fdecda509ac2de5656b94629c0855d5388
|
leogr
approved these changes
Sep 16, 2022
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: jasondellaluce, leogr, mstemm The full list of commands accepted by this bot can be found here. The pull request process is described here
Needs approval from an approver in each of these files:
Approvers can indicate their approval by writing |
mstemm
added a commit
that referenced
this pull request
Oct 25, 2022
The optimization in #2210 had a bug when the engine uses multiple sources at the same time--m_syscall_source is a pointer to an entry in the indexed vector m_sources, but if add_source is called multiple times, the vector is resized, which copies the structs but invalidates any pointer to the vector entries. So instead of caching m_syscall_source in add_source(), cache it in process_events(). m_sources won't change once processing events starts. Signed-off-by: Mark Stemm <mark.stemm@gmail.com>
poiana
pushed a commit
that referenced
this pull request
Oct 27, 2022
The optimization in #2210 had a bug when the engine uses multiple sources at the same time--m_syscall_source is a pointer to an entry in the indexed vector m_sources, but if add_source is called multiple times, the vector is resized, which copies the structs but invalidates any pointer to the vector entries. So instead of caching m_syscall_source in add_source(), cache it in process_events(). m_sources won't change once processing events starts. Signed-off-by: Mark Stemm <mark.stemm@gmail.com>
jasondellaluce
pushed a commit
to leogr/falco
that referenced
this pull request
Nov 17, 2022
The optimization in falcosecurity#2210 had a bug when the engine uses multiple sources at the same time--m_syscall_source is a pointer to an entry in the indexed vector m_sources, but if add_source is called multiple times, the vector is resized, which copies the structs but invalidates any pointer to the vector entries. So instead of caching m_syscall_source in add_source(), cache it in process_events(). m_sources won't change once processing events starts. Signed-off-by: Mark Stemm <mark.stemm@gmail.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
What type of PR is this?
/kind cleanup
Any specific area of the project related to this PR?
/area engine
What this PR does / why we need it:
Some small changes to rules matching that reduce cpu usage with high event volumes (> 1M syscalls/sec)
Which issue(s) this PR fixes:
Fixes #
Special notes for your reviewer:
Does this PR introduce a user-facing change?: