Skip to content

Roadmap stops keeping the state, and points at the issues that do - #109

Merged
Gerrrt merged 1 commit into
mainfrom
docs/roadmap-points-at-issues
Aug 21, 2026
Merged

Roadmap stops keeping the state, and points at the issues that do#109
Gerrrt merged 1 commit into
mainfrom
docs/roadmap-points-at-issues

Conversation

@Gerrrt

@Gerrrt Gerrrt commented Aug 21, 2026

Copy link
Copy Markdown
Owner

Follow-up to the 45 issues filed off #12.

Every open roadmap item now has an issue. Keeping the checkbox in
docs/roadmap.md as well means two places assert the same fact — and this file
has already lost that argument twice: it claimed the switch was not answering
SNMP while Prometheus held 280 series from it, and listed the git-history purge
as outstanding after it was done.

So the checkboxes go and the links stay. What the roadmap keeps is the thing an
issue tracker is bad at: the shape of the work and the order it wants doing in,
with the couplings visible — the firewall spare unblocks the restore rehearsal,
the playground waits on the main network, the eero integration waits on hardware
that has not been bought.

Nothing was summarised away. The trap in the MokerLink walk, the exact order
to bring the UPS back and why its silence must be deleted rather than left to
expire, ADR-0007's deferred isolation question for ifrit — all of it moved to
the issues verbatim before this file was trimmed.

Two things the reorganisation made obvious

  • A "Decided but not built" section. ADR-0007 and ADR-0008 between them
    describe a Windows domain, a second observability stack, nine services on
    hardware that does not exist, and two new firewall rules — and none of it
    appeared anywhere in this file.
  • A note that the stack does not watch itself, which is how two collection
    faults ran healthy and empty for half a day.

One item keeps its checkbox

"Rack the shelf switch" landed after the issues were filed and does not have one.
It stays as a checkbox and the file says so, rather than leaving it looking like
an oversight. Happy to file it if you want the set complete.

Scope

The README's pointer sentence is corrected here because this change is what made
it wrong. The stale top-items list beneath it is left alone — that is #75, per
the decision to file documentation drift rather than fix it.

Every one of the 29 issue links was checked to resolve.

🤖 Generated with Claude Code

Every open item now has an issue. Keeping the checkbox here as well means two
places assert the same fact, and this file has already lost that argument twice:
it said the switch was not answering SNMP while Prometheus held 280 series from
it, and listed the git-history purge as outstanding after it was done.

So the checkboxes go and the links stay. The roadmap keeps what an issue tracker
is bad at — the shape of the work and the order it wants doing in, with the
couplings visible: the firewall spare unblocks the restore rehearsal, the
playground waits on the main network being finished, the eero integration waits
on a machine that has not been bought.

Detail moved to the issues intact rather than being summarised away — the trap
in the MokerLink walk, the exact order to bring the UPS back and why the silence
must be deleted rather than left to expire, the isolation question ADR-0007
defers about ifrit.

Two additions this reorganisation made obvious:

- A "Decided but not built" section. ADR-0007 and ADR-0008 between them describe
  a Windows domain, a second observability stack, nine services on hardware that
  does not exist and two new firewall rules, and none of it appeared anywhere in
  this file. An accepted ADR with nothing tracking it is indistinguishable from a
  rejected one after six months.
- A note under Monitoring that the stack does not watch itself, which is how two
  collection faults ran healthy and empty for half a day.

"Rack the shelf switch" keeps its checkbox: it is the one item here without an
issue, and the file now says so rather than leaving it looking like an oversight.

The README's pointer is corrected in the same commit because this change is what
made it wrong. The stale top-items list beneath it is left alone — that is #75.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
@Gerrrt
Gerrrt merged commit f4e9262 into main Aug 21, 2026
3 checks passed
@Gerrrt
Gerrrt deleted the docs/roadmap-points-at-issues branch August 21, 2026 16:42
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant