Mark host interfaces as unused #35

Merged

jared merged 1 commits from unused-interfaces into main 2026-11-11 17:50:14 -05:00

Conversation 0 Commits 1 Files Changed 9

+219 -14

jared commented 2026-11-11 17:20:07 -05:00

Owner

Copy Link

Copy Source

Why

Some ports are never meant to be up: compute-storage-01's nic1-nic3, and large1's spare NIC and USB adapter. They kept their hosts "degraded", and they were behind tonight's false alerts (#31, reverted in #33).

What changes

Device page. A down interface gets a Mark unused button (admins only) with an optional reason, and In use again undoes it. Unused rows are greyed out with a dashed dot and an "unused" badge; the reason shows on hover. The header count leaves them out, for example "3 of 3 up · 1 unused".

Monitor. An unused interface:

• raises no link-down alert;
• can't count toward the cluster-wide alert;
• doesn't count toward the host's status;
• shows as unused in the snapshot, so the dashboard host cards and the rack drawings grey it out too.

The list is reloaded every cycle.

Marking closes any link-down alert open on that interface, without a recovery note, because nothing recovered.

API. PUT / DELETE /api/device/<host>/interfaces/<iface>/unused requires admin rights. PUT takes a JSON body {reason}. Interface names are validated. /api/device/<host> returns unused_interfaces.

Schema. A new table, unused_interfaces, is created with the cutsheet tables.

Candidates on production now

These are all the down interfaces:

• compute-storage-01: nic1, nic2, nic3
• large1: enp36s0, enx66630b4cce4a
• micro1: enp3s0. Leave this one: it's meant to be up and is waiting on a cable (ticket 666018103).

Testing

• Automated: 378 tests pass, 8 new:
  • unused interfaces never alert
  • they don't degrade the host, and show as unused in the snapshot
  • API mark and clear, admin, JSON and name checks, and the device API listing
• MariaDB:
  • marking closes an open alert quietly, with recovery_notified set so no note is queued
  • marking again updates the reason
  • clear is idempotent
• Playwright: mark, then the row turns unused and the count updates, then undo; no JS errors.
• Lint: flake8 is clean and ESLint reports 0 errors.

🤖 Generated with Claude Code

## Why Some ports are never meant to be up: compute-storage-01's `nic1`-`nic3`, and large1's spare NIC and USB adapter. They kept their hosts "degraded", and they were behind tonight's false alerts (#31, reverted in #33). ## What changes **Device page.** A down interface gets a **Mark unused** button (admins only) with an optional reason, and **In use again** undoes it. Unused rows are greyed out with a dashed dot and an "unused" badge; the reason shows on hover. The header count leaves them out, for example "3 of 3 up · 1 unused". **Monitor.** An unused interface: - raises no link-down alert; - can't count toward the cluster-wide alert; - doesn't count toward the host's status; - shows as `unused` in the snapshot, so the dashboard host cards and the rack drawings grey it out too. The list is reloaded every cycle. **Marking** closes any link-down alert open on that interface, without a recovery note, because nothing recovered. **API.** `PUT` / `DELETE /api/device/<host>/interfaces/<iface>/unused` requires admin rights. `PUT` takes a JSON body `{reason}`. Interface names are validated. `/api/device/<host>` returns `unused_interfaces`. **Schema.** A new table, `unused_interfaces`, is created with the cutsheet tables. ## Candidates on production now These are all the down interfaces: - compute-storage-01: `nic1`, `nic2`, `nic3` - large1: `enp36s0`, `enx66630b4cce4a` - micro1: `enp3s0`. Leave this one: it's meant to be up and is waiting on a cable (ticket 666018103). ## Testing - **Automated:** 378 tests pass, 8 new: - unused interfaces never alert - they don't degrade the host, and show as `unused` in the snapshot - API mark and clear, admin, JSON and name checks, and the device API listing - **MariaDB:** - marking closes an open alert quietly, with `recovery_notified` set so no note is queued - marking again updates the reason - clear is idempotent - **Playwright:** mark, then the row turns unused and the count updates, then undo; no JS errors. - **Lint:** flake8 is clean and ESLint reports 0 errors.  🤖 Generated with [Claude Code](https://claude.com/claude-code)

jared added 1 commit 2026-11-11 17:20:07 -05:00

Mark host interfaces as unused ... 3834546f38

Ports that are never meant to be up (compute-storage-01 nic1-3, large1's
spare NIC and USB adapter) kept their hosts "degraded" and were the source
of tonight's false alerts. An admin can now mark an interface unused from
the device page, with an optional reason. Unused interfaces raise no
link-down alerts, don't count toward the cluster-wide alert or the host's
status, and show greyed as "unused" on the dashboard, device page and rack.
Marking one closes any open link-down alert on it without a recovery note.

New table unused_interfaces, created with the cutsheet tables.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

jared scheduled this pull request to auto merge when all checks succeed 2026-11-11 17:50:11 -05:00

jared merged commit 6b03fe7b44 into main 2026-11-11 17:50:14 -05:00

Sign in to join this conversation.

Reviewers

No Reviewers

Labels

No items

No labels

Milestone

No items

No Milestone

Projects

Clear projects

No projects

Assignees

Clear assignees

cole (Cole Keller) jared (Jared Vititoe) nathan (Nathan Vititoe)

No Assignees

1 Participants

Notifications

Subscribe

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: LotusGuild/gandalf#35