carlos / platform Public

status-instances

Closes discussion #11: carlos status --instances prints one row per awake instance across the fleet — box, host, account, app, CPU, memory — through the console, needing no AWS credentials. --account <sqid> narrows it.

Nothing new is measured. The metrics loop already writes per-unit cpu and mem every 60s, LoadMetrics already sweeps them for the browser page, and unitOwner already joins a sampled cgroup to its route's account. All that was missing was a door.

Worth a reviewer's eye:

  • The door is registered with an EMPTY svcrole.Set — shut to every service credential, and added deliberately to closedDoors in svcroutes_test.go with the reason. The operator check is in the handler, because cliRoute's Set only narrows service principals and a person always passes it.
  • unitOwner's matching rule is split out as routeForUnit and shared rather than reimplemented, so this listing and the economics pass cannot come to disagree about whose a cgroup is.
  • A non-operator gets the byte-identical ghost 404, pinned in a test against a genuinely unknown route.

Not built: the "how long resident" column the issue asked for. Nothing in the tree publishes a process start time, so the only available answer would be "first seen in the 48h sweep" — under-reporting, and resetting on any gap. It needs a new field on the box's own sample; said so in the commit and in the issue rather than shipping a column that lies quietly.

Commits 0
Discussions 0 0 open
Tasks 0/0 done

Discussions

  • No discussions on this branch yet.

Tasks

  • No tasks yet.

Activity

  • No activity yet.