Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

quota: realtime usage update for metrics, logical/physical sizes, and data points #474

Draft
wants to merge 1 commit into
base: master
Choose a base branch
from

Conversation

bom-d-van
Copy link
Member

@bom-d-van bom-d-van commented Jun 28, 2022

Without this patch, it is possible for some namespaces to have a window of 1-3 carbonserver.quota-usage-report-frequency
to conduct quota enforcement penetration.

There are two cases that need to be addressed, dir nodes that are already annotated with proper quota config, nodes that
don't have quota config due to realtime insert via *CarbonserverListener.newMetricsChan.

For annotated nodes, trie index resovle it via realtime metric counter update in *trieIndex.insert.

For unannotated nodes, a new config are introduced, which is called transient-child-limit. With this config, during realtime
insert, trieIndex would annotate the new dir nodes of parent nodes that enables that config. This is not the most proper way
to address the issue, but a trade off of correctness, reliability, complexity, and performance. Re-iterating all quota configs
for every realtime insert is nice but expensive. However, as always, I could be wrong. But this seems to be working in my tests.

… data points

Without this patch, it is possible for some namespaces to have a window of 1-3 `carbonserver.quota-usage-report-frequency`
to conduct quota enforcement penetration.

There are two cases that need to be addressed, dir nodes that are already annotated with proper quota config, nodes that
don't have quota config due to realtime insert via *CarbonserverListener.newMetricsChan.

For annotated nodes, trie index resovle it via realtime metric counter update in *trieIndex.insert.

For unannotated nodes, a new config are introduced, which is called transient-child-limit. With this config, during realtime
insert, trieIndex would annotate the new dir nodes of parent nodes that enables that config. This is not the most proper way
to address the issue, but a trade off of correctness, reliability, complexity, and performance. Re-iterating all quota configs
for every realtime insert is nice but expensive. However, as always, I could be wrong. But this seems to be working in my tests.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

1 participant