Skip to content

S5b: hopin is routed through Traefik and isolated, trusting only Traefik's forwarded addresses #117

Closed

Description

@FlexOr2

hopin is reachable through Traefik at hopin.hallucinai.de from namespace cloudflared, isolated by a NetworkPolicy that admits only Traefik, and trusts only Traefik's forwarded addresses; the WebSocket timeout is decided by a driven cell. Proven in the job cluster on pull requests and on main.

Caller

Ruled

#46's section Ruled holds for this slice, each bullet with its own source (R1, R3a, R3b, Q6, Q8, Q9, #1's invariant 1); it is not copied again. #82's body quotes it verbatim as of 07.11.2026.

Decisions

#46's decisions 8, 9 and 10 hold. They live on #46 and are not copied here. #103 has landed Traefik's side of decision 9: its web admits only namespace cloudflared and trusts 10.42.0.0/16.

Done when

Copied verbatim from #46's Done when, except S8, cut to its route half: apps/ landed with S5a, and R1, R3a and R3b are ruled.

  • S4 (S5b) From namespace cloudflared, GET /health with Host hopin.hallucinai.de through Traefik's web gets 200, and an unknown host still gets 404. A Pod in another namespace gets no answer from hopin's Service.
  • S5 (S5b) HOPIN_TRUSTED_PROXIES lands with the NetworkPolicy of S4, in one change. With it, 40 sign-in starts from one forwarded address make the 41st get 429, while the next start from another forwarded address does not. Every Traefik pod's address lies in the setting.
  • S6 (S5b) H12 is reported with both connections' close times, and readTimeout stays at the default or is 0 as its result row in decision 10 says.
  • S8 (route half) AGENTS.md names the route, and quality hygiene . prints hygiene: clean.

Scenario matrix

Copied verbatim from #46's matrix. H12 is the explorative tester's, on a scratch pull request; H14 is read at landing.

Cell What is driven Expected observable Proven by
H8 S4, pull request /health through Traefik with hopin's host gets 200; an unknown host gets 404 CI
H9 S4, pull request a Pod in another namespace gets no answer on hopin:8080 within its timeout CI
H10 S5, pull request 40 /auth/start with X-Forwarded-For: 192.0.2.1, then the 41st gets 429; then one with 192.0.2.2 does not CI
H11 S5, pull request each Traefik pod's IP inside a CIDR of hopin's HOPIN_TRUSTED_PROXIES, both read from the cluster; the setting and the NetworkPolicy in the same diff CI and diff review
H12 S6, scratch: whoami's /echo behind a scratch Ingress; from namespace cloudflared, two websocat connections held 120s at the default readTimeout, one idle and one with --ping-interval 20; again at 0 if decision 10's table says so each connection's close time or "held"; the table row they match and the value it decides driven
H14 S1 to S6, push to main cluster green with H1 to H3 and H8 to H11's observables CI, read at landing

Files

  • apps/hopin/hopin.yaml: the Ingress, the NetworkPolicy and HOPIN_TRUSTED_PROXIES.
  • .github/workflows/ci.yml, job cluster: the route, NetworkPolicy, rate-limit and relationship cells, whose client Pods run in namespace cloudflared.
  • AGENTS.md: the cluster line on the route.
  • infrastructure/traefik/values.yaml and its generated rendered.yaml, only if H12 cuts the pinging connection: readTimeout: 0 on web, regenerated with mise run charts:render.

Neighbours

version = 1
now = "Cut from #46 (decision 11's S5b) on 07.11.2026. #103 (S4b) has landed; S5a (#82, PR #96) is ready to land after its merge of main."
next = "Build once #82 has landed, from main."
done_when = "From namespace cloudflared, /health with hopin's host through Traefik's web gets 200 and an unknown host 404, a Pod in another namespace gets no answer from hopin's Service, HOPIN_TRUSTED_PROXIES lands with the NetworkPolicy and makes the 41st sign-in start from one forwarded address get 429 but not the next from another, every Traefik pod's address lies in the setting, H12 is reported and readTimeout follows its row, AGENTS.md names the route, quality hygiene . prints hygiene: clean, and H8 to H12 and H14 are reported."
scope = ["apps/hopin/hopin.yaml", ".github/workflows/ci.yml", "AGENTS.md", "infrastructure/traefik/values.yaml", "infrastructure/traefik/rendered.yaml"]
size = "M"
whole = "S5b does not split: the route, the NetworkPolicy and the trusted proxies are one trust boundary (decision 9), and Traefik's timeout follows only from H12 on that route."

Activity

  1. added theissue type on Nov 7, 2026
  2. FlexOr2 commented on Nov 7, 2026

    @FlexOr2
    MemberAuthor

    Journal, lane PR #118.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions