refactor(userspace/engine): strengthen and document thread-safety guarantees of falco_engine::process_event #2082
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
jasondellaluce
changed the title
jasondellaluce
force-pushed
the
ref/statsmanager-tsafe
branch
from
9848b67 to
706f849
Compare
|
/milestone 0.33.0 |
jasondellaluce
commented
Jun 21, 2022
| std::vector<uint64_t> m_by_priority; | ||
| std::vector<uint64_t> m_by_rule_id; | ||
| atomic<uint64_t> m_total; | ||
| std::vector<std::unique_ptr<atomic<uint64_t>>> m_by_priority; |
There was a problem hiding this comment.
Note: the usage of unique_ptr is required because atomic<> does not have copy nor move constructors, which makes them unusable in vectors as simple values.
jasondellaluce
changed the title
15 tasks
jasondellaluce
force-pushed
the
ref/statsmanager-tsafe
branch
2 times, most recently
from
ff30d60 to
343ceac
Compare
…nt method Signed-off-by: Jason Dellaluce <[email protected]>
…e::process_event Signed-off-by: Jason Dellaluce <[email protected]>
Signed-off-by: Jason Dellaluce <[email protected]>
Signed-off-by: Jason Dellaluce <[email protected]>
jasondellaluce
force-pushed
the
ref/statsmanager-tsafe
branch
from
343ceac to
1c873ce
Compare
|
LGTM label has been added. Git tree hash: 722c6bad23695db9ab69de98bb4de599ab91fbdc
|
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: FedeDP, jasondellaluce 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 |
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Outdated suggestions cannot be applied.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 feature
Any specific area of the project related to this PR?
/area engine
What this PR does / why we need it:
This PR is focused on guaranteeing (and documenting) that concurrent access to
falco_engine::process_event()is safe for different event sources.Since the new plugin system got implemented, and due to the recent refactoring of the Falco engine for Falco v0.32, the engine is not internally split by event sources. This means that the engine contains multiple rulesets, each mapped 1-1 with an event source. Accordingly, it is possible to assume that
falco_engine::process_event()is now thread-safe with some assumptions:source_idxsource_idxacross subsequent calls tofalco_engine::process_event()Considering the above, the only synchronization point is
stats_manager::on_event(), which has been modified to be thread-safe and lock-free.Which issue(s) this PR fixes:
Special notes for your reviewer:
Thread safety is not a requirement for Falco, but future developments such as the ones described in #2074 could require it. Among the many things that can be useful to make thread safe, the most crucial one is the rule-event matching routine, of which
falco_engine::process_event()is entrypoint. With this implementation, it is possible to consider processing events coming from different event sources in different parallel threads.Assuming a single-threaded scenario, the changes involved in
stats_manager::on_event()are a no-op in both expected behavior and performance. This is true becausestats_manager::on_event()is invoked when an event matches a given rule exclusively, which happens with less frequency than the production of events by some orders of magnitude. As such, we can consider these changes as non-noticeable for single-thread scenarios (which is what we have right now). Moreover, when events are matched with a rule we then have to queue them in the Falco output engine, which is still an asynchronous sequential synchronization point.Does this PR introduce a user-facing change?: