Skip to content
This repository has been archived by the owner on Nov 15, 2023. It is now read-only.

Add buckets on lower end of distribution to network bridge latency metrics #6359

Merged
merged 6 commits into from
Dec 2, 2022
Merged

Add buckets on lower end of distribution to network bridge latency metrics #6359

Changes from 4 commits
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
9a409d4
Add two more buckets on lower end of distribution
bredamatt Nov 28, 2022
8c3fc70
Merge branch 'master' of github.com:paritytech/polkadot into bredamat…
bredamatt Nov 29, 2022
11e0532
add even smaller buckets
bredamatt Nov 29, 2022
ca91997
Merge branch 'master' of github.com:paritytech/polkadot into bredamat…
bredamatt Nov 30, 2022
c8f10a4
cargo fmt
bredamatt Dec 2, 2022
73f2bf6
Merge branch 'master' of github.com:paritytech/polkadot into bredamat…
bredamatt Dec 2, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion node/network/statement-distribution/src/metrics.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -18,7 +18,7 @@ use polkadot_node_subsystem_util::metrics::{self, prometheus};

/// Buckets more suitable for checking the typical latency values
const HISTOGRAM_LATENCY_BUCKETS: &[f64] =
&[0.05, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.75, 0.9, 1.0, 1.2, 1.5, 1.75];
&[0.000025, 0.00005, 0.000075, 0.0001, 0.0003125, 0.000625, 0.00125, 0.0025, 0.005, 0.01, 0.025, 0.05, 0.1];
Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Locally these buckets don't make a big difference, but I am interested in seeing if they do in a more realistic scenario.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

For some realistic values I would look at some historical data from Versi. Even if we don't have these buckets, the sum and count of samples allow us to compute some average value which could be used to estimated some good buckets.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Good idea!


#[derive(Clone)]
struct MetricsInner {
Expand Down