feat(backend): reconcile running bottles' CA on gateway bring-up (0081)
tracker-policy-pr / check-pr (pull_request) Successful in 15s
lint / lint (push) Successful in 1m15s
test / unit (pull_request) Successful in 51s
test / integration-docker (pull_request) Failing after 1m35s
test / coverage (pull_request) Has been skipped

Add `attach_bottled_agents_to_gateway()` as an `@abc.abstractmethod` on
`BottleBackend` and implement the first reconcile step across all three
backends: on a gateway cold boot, replace every already-running agent's
trusted CA with the freshly-minted gateway CA. Reconciling running bottles
is a backend responsibility (only the backend can enumerate its agents and
reach them — firecracker over SSH, docker/macOS over exec/cp), so making it
an abstract method keeps the fix cross-backend by construction.

The gateway rootfs is ephemeral, so a rebuild mints a new CA that every
running bottle distrusts (SSL verification fails — #510). This reconcile is
what distributes the fresh CA, so a routine gateway rebuild now doubles as a
free CA rotation. Per-bottle steps are best-effort: one unreachable or
malformed bottle is logged and skipped, never blocking the others.

Trigger — `Gateway.connect_to_orchestrator` now returns a cold-boot bool
(True when it actually (re)brought the gateway up, False when a healthy
current gateway was left untouched); each infra `ensure_running` gates the
reconcile on it, so it fires exactly on cold boot and never on an adopt.

Git-gate re-provision and egress-token restore fold into this same reconcile
on the same trigger in follow-up PRs (#516).

Refs #516. Addresses #510.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
This commit is contained in:
2026-07-26 23:54:27 +00:00
parent 4434752db4
commit de80aafb1f
21 changed files with 488 additions and 15 deletions
+24
View File
@@ -29,10 +29,20 @@ class TestDockerInfraService(unittest.TestCase):
self.orch.gateway_url.return_value = f"http://{ORCHESTRATOR_NAME}:8099"
self.orch.mint_gateway_token.return_value = "gw.jwt"
self.gw = MagicMock()
# Default to a cold boot (gateway (re)created) so the reconcile branch
# runs; the actual reconcile is stubbed so these composition tests stay
# isolated from docker.
self.gw.connect_to_orchestrator.return_value = True
for name, mock in (("orchestrator", self.orch), ("gateway", self.gw)):
p = patch.object(self.svc, name, return_value=mock)
p.start()
self.addCleanup(p.stop)
ap = patch(
"bot_bottle.backend.docker.backend.DockerBottleBackend"
".attach_bottled_agents_to_gateway"
)
self.attach = ap.start()
self.addCleanup(ap.stop)
def test_url_delegates_to_the_orchestrator(self) -> None:
self.assertEqual("http://127.0.0.1:8099", self.svc.url())
@@ -58,6 +68,20 @@ class TestDockerInfraService(unittest.TestCase):
f"http://{ORCHESTRATOR_NAME}:8099", "gw.jwt")
self.orch.mint_gateway_token.assert_called_once()
def test_cold_boot_reconciles_running_bottles(self) -> None:
# A (re)created gateway (connect returns True) triggers the bring-up
# reconcile so running bottles re-trust the fresh gateway (PRD 0081).
self.gw.connect_to_orchestrator.return_value = True
self.svc.ensure_running()
self.attach.assert_called_once_with()
def test_no_reconcile_when_gateway_left_untouched(self) -> None:
# A healthy current gateway (connect returns False) is not a cold boot,
# so there is nothing to reconcile.
self.gw.connect_to_orchestrator.return_value = False
self.svc.ensure_running()
self.attach.assert_not_called()
def test_stop_removes_both_containers(self) -> None:
with patch(_RUN) as run:
self.svc.stop()