perf: compare service discovery nodes by addresss#12258
Merged
Conversation
nic-6443
force-pushed
the
nic/compare-nodes-by-addr
branch
from
May 26, 2025 09:30
57e2a81 to
6db2724
Compare
Signed-off-by: Nic <[email protected]>
nic-6443
force-pushed
the
nic/compare-nodes-by-addr
branch
from
May 26, 2025 09:31
6db2724 to
7b153c5
Compare
nic-6443
requested review from
AlinsRan,
Revolyssup,
bzp2010,
membphis,
nic-chen and
ronething
May 27, 2025 01:01
nic-chen
approved these changes
May 27, 2025
AlinsRan
approved these changes
May 27, 2025
membphis
approved these changes
May 29, 2025
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Description
When the number of nodes discovered by the service is large, the comparison based on node content in the
compare_upstream_nodefunction will be very CPU-intensive. Since every request goes throughcompare_upstream_node, this results in much higher CPU consumption for upstreams based on service discovery compared to regular upstreams.The current PR optimization approach is to first compare the memory addresses of
nodestables incompare_upstream_node. If the addresses are identical, it can quickly determine that twonodesare identical.For this optimization to work, the following conditions must be met:
service_namecorresponding to itsnodes, a new table should be used to store data. I have confirmed that all existing service discoveries meet this condition.fill_nodesfunction, modifications to upstream data should not be made by cloning an old nodes table but rather by reusing the same table.Benchmark Result
Before
1 Nodes
Upstream Nodes
Service Discovery
100 Nodes
Upstream Nodes
Service Discovery
After
1 Nodes
Upstream Nodes
Service Discovery
100 Nodes
Upstream Nodes
Service Discovery
Summary
Checklist