pulsar
pulsar copied to clipboard
[improve][offload]PIP-211: Introduce Offload throttling, P1
PIP: https://github.com/apache/pulsar/issues/18004
Motivation
See https://github.com/apache/pulsar/issues/18004
Modifications
- Move AsyncTokenBucket related classes from pulsar-broker to pulsar-common
- Add OffloadLedgerHandle.java
- Pass OffloadLedgerHandle's instance to LedgerOffloader instead of LedgerHandle.
Verifying this change
- [ ] Make sure that the change passes the CI checks.
(Please pick either of the following options)
This change is a trivial rework / code cleanup without any test coverage.
(or)
This change is already covered by existing tests, such as (please describe tests).
(or)
This change added tests and can be verified as follows:
(example:)
- Added integration tests for end-to-end deployment with large payloads (10MB)
- Extended integration test for recovery after broker failure
Does this pull request potentially affect one of the following parts:
If the box was checked, please highlight the changes
- [ ] Dependencies (add or upgrade a dependency)
- [ ] The public API
- [ ] The schema
- [ ] The default values of configurations
- [ ] The threading model
- [ ] The binary protocol
- [ ] The REST endpoints
- [ ] The admin CLI options
- [ ] The metrics
- [ ] Anything that affects deployment
Documentation
- [ ]
doc
- [x]
doc-required
- [ ]
doc-not-needed
- [ ]
doc-complete
Matching PR in forked repository
PR in forked repository:
Origin PR: https://github.com/apache/pulsar/pull/17485
I think that AsyncTokenBucket should be used for rate limiting instead of implementing a separate solution. It was introduced in PIP-322.
@lhotari TimeWindow is not introduced by this PR, it exists a long time. Just reuse it here. If you think AsyncTokenBucket is better, I can fix this.
@lhotari I will move AsyncTokenBucket related class from pulsar-broker to pulsar-common, is it OK?
I think that AsyncTokenBucket should be used for rate limiting instead of implementing a separate solution. It was introduced in PIP-322.
@lhotari TimeWindow is not introduced by this PR, it exists a long time. Just reuse it here. If you think AsyncTokenBucket is better, I can fix this.
Yes, I think so. It's worth checking if you could implemented the rate limiting in a similar way as it is in other Pulsar rate limiters that were refactored in PIP-322 implementation.
Codecov Report
Attention: Patch coverage is 83.05085%
with 10 lines
in your changes are missing coverage. Please review.
Project coverage is 73.71%. Comparing base (
bbc6224
) to head (ada5b82
). Report is 107 commits behind head on master.
Additional details and impacted files
@@ Coverage Diff @@
## master #22385 +/- ##
============================================
+ Coverage 73.57% 73.71% +0.13%
+ Complexity 32624 32404 -220
============================================
Files 1877 1886 +9
Lines 139502 139926 +424
Branches 15299 15326 +27
============================================
+ Hits 102638 103141 +503
+ Misses 28908 28810 -98
- Partials 7956 7975 +19
Flag | Coverage Δ | |
---|---|---|
inttests | 27.03% <11.86%> (+2.45%) |
:arrow_up: |
systests | 24.48% <52.54%> (+0.16%) |
:arrow_up: |
unittests | 72.98% <79.66%> (+0.14%) |
:arrow_up: |
Flags with carried forward coverage won't be shown. Click here to find out more.
Files | Coverage Δ | |
---|---|---|
...apache/bookkeeper/mledger/ManagedLedgerConfig.java | 96.38% <100.00%> (+0.08%) |
:arrow_up: |
...che/bookkeeper/mledger/impl/ManagedLedgerImpl.java | 80.95% <100.00%> (+0.29%) |
:arrow_up: |
...org/apache/pulsar/broker/ServiceConfiguration.java | 99.39% <100.00%> (+<0.01%) |
:arrow_up: |
...n/java/org/apache/pulsar/broker/PulsarService.java | 82.62% <ø> (+0.25%) |
:arrow_up: |
...ker/resourcegroup/ResourceGroupPublishLimiter.java | 86.95% <ø> (ø) |
|
...rg/apache/pulsar/broker/service/BrokerService.java | 80.96% <100.00%> (+0.18%) |
:arrow_up: |
.../pulsar/broker/service/PublishRateLimiterImpl.java | 86.15% <ø> (ø) |
|
...broker/service/persistent/DispatchRateLimiter.java | 78.63% <ø> (ø) |
|
...roker/service/persistent/SubscribeRateLimiter.java | 54.54% <ø> (ø) |
|
...ava/org/apache/pulsar/broker/stats/TimeWindow.java | 66.66% <ø> (ø) |
|
... and 8 more |