-
Notifications
You must be signed in to change notification settings - Fork 55
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
feat: drop msgs to be relayed waiting for too long in the queue #1017
base: master
Are you sure you want to change the base?
Conversation
800dd18
to
61a1300
Compare
Codecov ReportAttention:
Additional details and impacted files@@ Coverage Diff @@
## forward-msgs-non-priority-v2 #1017 +/- ##
================================================================
- Coverage 83.27% 83.24% -0.03%
================================================================
Files 91 91
Lines 15428 15438 +10
================================================================
+ Hits 12847 12852 +5
- Misses 2581 2586 +5
|
5006b79
to
b7b0e5d
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
At first glance it LGTM.
Maybe it'd be good to take into account Ludovic's opinion, given he's more familiar with the repo?
ed4cd60
to
5bba0f5
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks! Looks good.
Is there a reference as to why 1s was chosen as the default.
How do other implementations / the spec handle this?
If you did not do this yet, I suggest first testing the base PR in a testnet, followed by another run with this PR merged into the base PR forward-msgs-non-priority-v2
@kaiserd I've been testing all PRs running a local Nimbus client on Holesky. Everything seems fine. The only implementation with this feature I'm aware of is rust-libp2p and last time I checked they were testing it on a branch sigp/rust-libp2p#555. They use 500ms as the default value. I tried it first but I saw a large number of messages being dropped. I believe this value is network/application-specific and could be hard to guess a default value. 1s was just an initial attempt, I welcome feedback here. |
9fe2ec6
to
5b48aa4
Compare
Those are the other features they implemented sigp/lighthouse#4918 (comment) |
2c37aa4
to
311893e
Compare
311893e
to
38c0af4
Compare
826ce1b
to
07cab43
Compare
f7ae9fb
to
136e88a
Compare
136e88a
to
a04f8d2
Compare
eed49c6
to
8861515
Compare
8861515
to
6484d3f
Compare
We need further investigation here. |
This PR adds the possibility of configuring the maximum time a message can stay in the non-priority queue and still be considered relevant to be sent to the respective remote peer. If the message has been waiting for this duration or longer at the time is dequeued it will be discarded and not sent.