Skip to content

Sealane v1.4.0

Faster mesh views and a Services list that tells you where the routing is

Highlights

  • One concurrent snapshot per request. The traffic flow, mesh analysis, service view and route check used to issue their Kubernetes lists one after another and fetch Endpoints once per backend Service. They now load one snapshot concurrently and read Endpoints once; each view requests only the kinds it uses. On a cluster with ~240 pods and ~100 Services and an API server answering in ~100 ms, the sequential version took about 3 s per page; see Architecture.
  • Client rate limit raised. Sealane now sets client-go to 50 requests/s (burst 100) instead of the 5/10 default shared by every user, so two people opening pages at once no longer queue behind each other on the client side.
  • Services list with mesh counts. /services shows, per Service, how many routes (VirtualServices and HTTPRoutes), destination rules and security policies apply to it, and how many analysis findings concern it — sortable, with findings as a red badge. The numbers come from the same code as the Service page, so they always agree.

Upgrade notes

  • No breaking changes; existing sessions stay valid. The chart is unchanged apart from the version.
  • helm upgrade sealane oci://ghcr.io/wonjune95/charts/sealane --version 1.4.0 -n <namespace> --reset-then-reuse-values

Install

helm install sealane oci://ghcr.io/wonjune95/charts/sealane --version 1.4.0 -n sealane --create-namespace
# or
kubectl apply -f https://github.com/wonjune95/sealane/releases/download/v1.4.0/install.yaml
Image ghcr.io/wonjune95/sealane:1.4.0 (linux/amd64, linux/arm64) and the chart are signed with cosign (keyless); the SBOM is attached below.

Full Changelog: https://github.com/wonjune95/sealane/compare/v1.3.2...v1.4.0