docs: record sentinel frpc active-root conflict
This commit is contained in:
@@ -109,6 +109,8 @@ Classify a monitor page issue in this order:
|
||||
|
||||
Do not treat public root/CSS/JS 200 as dashboard success. Browser console and DOM render evidence are required.
|
||||
|
||||
When a Pod and its ClusterIP Service return the expected monitor-web asset but `https://monitor.pikapython.com/monitor-web/assets/...` still returns an old file, check the public FRP hop before changing UI code again. On D518 this has happened when a stale legacy `hwlab-web-probe-sentinel-frpc` Argo app still owned PK01 remote port `22093`, while the active `hwlab-web-probe-sentinel-fake-echo-frpc` logged `start error: port already used`; Caddy then served the old runner even though the new Pod and runtime image were healthy. Retire the stale Argo/frpc owner through a controlled YAML/CLI path and then rerun `dashboard verify`.
|
||||
|
||||
## Rollout Closeout
|
||||
|
||||
For a Web sentinel fix, closeout needs four independent evidence surfaces:
|
||||
|
||||
Reference in New Issue
Block a user