Repository navigation
Conversation
Tacklebox
reviewed
Aug 20, 2026
biscout42
commented
Aug 21, 2026
christos68k
reviewed
Aug 26, 2026
biscout42
force-pushed
the
fix-container-link
branch
from
September 16, 2026 11:38
77013a6 to
2bd0f78
Compare
biscout42
force-pushed
the
fix-container-link
branch
from
September 16, 2026 12:12
2bd0f78 to
f2cd116
Compare
biscout42
force-pushed
the
fix-container-link
branch
from
September 16, 2026 12:49
f2cd116 to
f8d799b
Compare
biscout42
force-pushed
the
fix-container-link
branch
from
September 16, 2026 13:07
f8d799b to
2e49209
Compare
biscout42
force-pushed
the
fix-container-link
branch
3 times, most recently
from
September 16, 2026 16:05
21e8cce to
01e181d
Compare
biscout42
force-pushed
the
fix-container-link
branch
3 times, most recently
from
September 24, 2026 11:29
38628b7 to
736b875
Compare
biscout42
force-pushed
the
fix-container-link
branch
from
September 24, 2026 11:59
736b875 to
89b0fdd
Compare
biscout42
force-pushed
the
fix-container-link
branch
from
September 25, 2026 08:29
89b0fdd to
96de704
Compare
Avoid using the lazy public lookup in internal event assembly.
Co-authored-by: Christos Kalkanis <christos.kalkanis@elastic.co>
…ue ops Link existing processes at queue open only when a kube talker is present. Without it the container tree is empty and the pass does nothing. Return early from quark_queue_populate when queue_ops is NULL. This removes the stub ops from the C and Go tests. Move the Go test queue helper into quark.go and remove the quarktest build tag. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
…backend-less queue Do not unlink a process from its container when the new cgroup does not name a different container. Do not call update_stats when the queue has no backend. Use process_cache_get in the mprotect event. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Install a queue_ops table that does nothing instead of a NULL pointer. Remove the NULL checks from populate and get_stats. Use the new function in all C and Go test fixtures. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
…ontainer Put the flag in the hole after pid. This saves 8 bytes per cached process. Replace three copies of the container unlink code with one helper. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
biscout42
force-pushed
the
fix-container-link
branch
from
September 25, 2026 17:17
96de704 to
c45e3f3
Compare
Member
|
I opened a subsequent PR (set to merge into this one) to address a race condition / ordering issue between kube-talker and API when creating a container. See: #438 |
This was referenced Sep 28, 2026
This branch has not been deployed
This file contains hidden or 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
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.
As we link container data from an external process there is no need to check kube talker.
That is a bit complex, because adds parsing cgroup to all events. I have added guards to reduce the performance footprint for non-k8s use case. It should behave like this:
Implications:
Imho the last point is acceptable because no events meaning no messages to deliver from a queue.
TODO:
quarktesttag. In this PR libquark exports quark_queue_init_trees in the header's "exported for testing only" section. With a one-line null guard on queue_ops in quark_queue_populate, matching the guard already in quark_queue_close, the stub ops become unnecessary. The Go helper then shrinks to a few lines that call calloc, call the init function, and set two fields, with no C preamble at all. It could sit unexported at the bottom of quark.go. That is test-only code in a production file, but it mirrors the precedent libquark already sets in its header. Precisely the same change also collapses the three hand-rolled C and Go fixtures, which is the fixture-duplication finding two items down the list.