Make the runner egress proxy always-on now that the allowlist is gone #784
Labels
No labels
burndown-2026-06
autonomy
async-consult
autonomy
epic
autonomy
headless
autonomy
live-collab
coherence-core
priority
P0
priority
P1
priority
P2
priority
P3
priority
P4
qa-fixture
role/ai
role/creator
role/design
role/director
role/engineer
role/exec
role/human
role/ops
role/qa
No milestone
No project
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set
Reference
coilyco-flight-deck/infrastructure#784
Loading…
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
a2f88cfdropped the destination allowlist on 2026-08-11(
config-revision: 2026-08-11-open-destinations). The current config ishttp_access allow CONNECT SSL_ports, so Squid tunnels CONNECT to any host on443.
That removes the reason the proxy was opt-in. Deny-by-default was what made
global proxying dangerous, because every unlisted destination would have broken.
Fix
Set
HTTP_PROXY,HTTPS_PROXY, andNO_PROXYin the runnerenvs:blocksbeside the existing
FORGEJO_EGRESS_PROXY. One change here replaces theper-repository wrappers.
Constraints
NO_PROXYmust cover Forgejo, or a checkout deadlocks against its own ingress.NO_PROXYmust cover the in-cluster registry. Squid allows CONNECT on 443only, and the DinD daemons carry
--insecure-registry=192.168.0.194:30500,which is plain HTTP.
Relationship to the MTU issue
This is a workaround for the MTU defect, not a fix. If the MTU issue lands
first, this becomes optional rather than necessary.