Reconcile: adopt initialization's consolidated HQ as canonical, re-home this session's new work #24

Merged
jschoubben merged 177 commits from reconcile-init-into-main into main 2026-09-05 10:27:11 +00:00
9 changed files with 98 additions and 42 deletions
Showing only changes of commit e1f4c7d9e0 - Show all commits
+7 -5
View File
@@ -114,12 +114,14 @@ and it runs on no node at all.
Anatomy makes attractive names and poor boundaries. Name the thing the domain calls it.
### Group by domain, not by single function
### Things that change together share an authority, not a package
A module is a purpose, not a piece of software. Four modules that together constitute "how a
node is reachable" and cannot be assigned, versioned or replaced as one thing are four
accidents, not four boundaries.
[ADR 0017](../02-DECISIONS/0017-modules-outside-the-core-are-grouped-by-domain.md)
When several modules always change together under one intent, name the **context** that decides
for them. Do not merge them into one module: they are delivered to different nodes, and a module
that must be assigned where half of it is unwanted is not a boundary either.
Coherence is a context. Delivery is a module. Relationships are edges, not folders.
[ADR 0054](../02-DECISIONS/0054-things-that-change-together-share-an-authority.md)
### Contexts integrate through the record, never through a shared schema
@@ -1,5 +1,6 @@
---
status: accepted
status: superseded
superseded-by: 02-DECISIONS/0056-the-authority-is-the-control-plane-not-a-database.md
date: 2026-04-02
deciders: jochen
reconstructed: true
@@ -47,9 +47,16 @@ beside `02-DECISIONS/`, holding different things.
**If a decision is worth recording, it is worth a record. If it is not worth a record, it is
not recorded.**
`02-DECISIONS` holds every decision. There is no ledger, no index file, and no central status
of any kind. The chronological view — decisions in the order they were taken — is *generated*
from record frontmatter, which is what the ledger was actually for.
`02-DECISIONS` holds every decision, and **there is no ledger** — no separate document in which
a decision is also summarised, ranked or tracked. The chronological view — decisions in the order
they were taken — is *generated* from record frontmatter, which is what the ledger was actually
for.
That generation is not this record's rule. It is
[ADR 0022](0022-status-lives-in-frontmatter.md), which already decides repository-wide that
status lives in frontmatter and every cross-cutting view is generated rather than written. This
record does not restate it — 0022's own words are *prose does not restate status; one place, and
two is one too many*, and an earlier version of this paragraph did exactly that.
Content that was only in the ledger was rehomed rather than dropped:
@@ -55,14 +55,35 @@ mesh.
## Decision
**The enrolment token carries three things**, and it is the only thing a joining node needs:
**The enrolment token carries four things**, and it is the only thing a joining node needs:
| | | |
|---|---|---|
| **where** | an **address**, not a name | there is no resolution yet, and this is why none is needed |
| **who** | the fingerprint of the identity to expect | what makes the mesh provable rather than assumed |
| **where** | the **broker's address**, not a name | a node dials the broker ([ADR 0001](0001-nodes-communicate-over-a-broker.md)); there is no resolution yet, and this is why none is needed |
| **what it is connecting to** | the fingerprint of the **broker's** certificate | so the node reaches the mesh's bus and not something answering in its place |
| **who it will believe** | the **control plane's** signing identity | what makes the mesh provable rather than assumed |
| **the right to join** | the one-time secret ADR 0039 already specifies | useless once used, useless after it expires |
### The endpoint and the authority are two identities, not one
This is worth separating because collapsing it is the easy mistake, and the collapsed version
silently fails to deliver what [ADR 0039](0039-the-link-is-the-security-boundary.md) asks for.
A node connects to the **broker** and takes instruction from the **control plane**, which sits
behind it. Pinning only the broker would make the control plane's authority *transitive* — the
node would believe a declaration because of where it arrived from. **A compromised broker could
then forge declarations**, and since the host applies whatever the link delivers, that is the
whole machine.
So the node verifies **the transport** and **each declaration** separately:
- the broker, by its certificate, at connect time;
- the control plane, by a **signature on the declaration itself**, every time.
Then 0039's *the control plane proves it is the mesh* holds against a hostile broker rather than
assuming a friendly one — which matters because the broker is the one component every node must
reach and the one most exposed.
The token is issued by the mesh for one enrolment and **carried out of band** — by the person
adopting the machine. That is what breaks both circles: its authenticity comes from the channel
it travelled, not from anything the node can check afterwards.
@@ -104,6 +125,13 @@ being a design statement and being a fact.
- **The address in the token can go stale.** If the control plane moves, unissued tokens point
somewhere wrong. Tokens are short-lived, which bounds it; moving the control plane is
[`06`](../03-DESIGN/01-to-be/06-the-control-plane.md)'s undesigned territory regardless.
- **Declarations must be signed, and that is a real requirement rather than a note.** The host
verifies a signature before applying anything, which adds a key to what it must carry and a
failure mode it must report legibly — *this declaration is not from the mesh I joined* is a
different condition from *this declaration is malformed*, and they must not read alike.
- **Rotating the control plane's signing identity is a fleet-wide operation**, because every node
holds the previous one. That is the cost of not trusting the broker, it is accepted, and it
needs a rollover that overlaps rather than a flag day.
- **A rejoining node is an ordinary case, not a special one.** A node that has lost its identity
gets a new token. There is no recovery path to design because there is no long-lived secret to
recover.
@@ -1,5 +1,5 @@
---
status: proposed
status: accepted
date: 2026-08-27
deciders: jochen
reconstructed: false
@@ -91,10 +91,16 @@ replacement keeps the observation and changes the instruction:
- **The instruction now matches the design.** An agent reading the constitution and an agent
reading 0044 reach the same answer, which they currently do not.
- **The republication is the point, not a formality.** Until the derived page is regenerated the
mesh is still governed by the old rule, and this record has changed nothing where it matters.
Verified by reading the rule back out of the published page
([ADR 0035](0035-a-picture-is-read-from-what-runs.md)), not by the publish reporting success.
- **Synced 2026-08-27**, playbook [05](../00-META/process/05-constitution-sync.md). The derived
page carries the new rule and §4 keeps its number. **Verified by reading back**, not by the
publish reporting success: the replacement text is present, and the old section's body — *scope
under measurement: grouping is evidenced for reachability* — returns nothing.
- **The sync found a drift the playbook exists to catch.** The published §4 and the source did not
say the same thing: the source spoke of *four accidents, not four boundaries*, the published
page of *one intent expressed four times*, and only the published page carried the scope
caveat. Same rule, two texts, already diverging — which is the exact failure
[ADR 0025](0025-hq-is-the-source-of-the-constitution.md) predicted and the reason the playbook
re-publishes the whole page rather than patching a section.
- **`context` becomes a word the constitution uses**, which raises the obvious next question —
*which contexts are there* — and that is
[ADR 0055](0055-the-control-plane-is-the-node-coordinating-contexts.md), not this record.
@@ -1,5 +1,5 @@
---
status: proposed
status: accepted
date: 2026-08-27
deciders: jochen
reconstructed: false
@@ -1,5 +1,5 @@
---
status: proposed
status: accepted
date: 2026-08-27
deciders: jochen
reconstructed: false
+29 -18
View File
@@ -8,6 +8,7 @@ decisions:
- 02-DECISIONS/0053-one-control-plane-and-no-failover.md
- 02-DECISIONS/0030-the-repository-structure.md
- 02-DECISIONS/0015-mesh-brokers-nodes-host-agents-think.md
- 02-DECISIONS/0055-the-control-plane-is-the-node-coordinating-contexts.md
---
# The control plane
@@ -42,22 +43,31 @@ catch it because the dependency direction is still correct.
## What is inside it
Ten contexts and one interface, from the skeleton
([research 006](../../01-RESEARCH/006-mesh-from-scratch/skeleton.md)):
**Seven contexts and one interface**
([ADR 0055](../../02-DECISIONS/0055-the-control-plane-is-the-node-coordinating-contexts.md)) —
each one earning its place by the test above rather than by being ours:
| | |
|---|---|
| **record** | the event log every other context integrates through |
| **inventory** | nodes, modules, assignments, versions |
| **config** | settings, secrets, and deriving them onto nodes |
| **connectivity** | overlay, resolution, exposure, filtering, certificates — **specified in full in [`08-connectivity.md`](08-connectivity.md)** |
| **provisioning** | resource grants between modules |
| **delivery** | source to artifact to node |
| **observability** | health, logs, metrics, alerts |
| **identity** | agents, humans, services, authorisation |
| **work** | tasks, workflows, runs |
| **knowledge** | memory, documents, retrieval |
| **api** | the one interface every surface speaks to |
| | | needs to know about more than one node because |
|---|---|---|
| **inventory** | nodes, modules, assignments, versions | that *is* the mesh-wide fact |
| **config** | settings, secrets, and deriving them onto nodes | it derives **onto nodes** |
| **connectivity** | overlay, resolution, exposure, filtering, certificates — **specified in full in [`08-connectivity.md`](08-connectivity.md)** | who peers with whom; which node is reachable |
| **provisioning** | resource grants between modules | consumer and provider may be on different nodes |
| **delivery** | source to artifact to node | it targets nodes |
| **observability** | health, logs, metrics, alerts | *unreachable for a week* is nobody else's to notice |
| **identity** | agents, humans, services, authorisation | credentials follow an agent's node bindings and modality |
| **api** | the one interface every surface speaks to | — it is an interface, not a context |
**What is deliberately not here.** `work`, `knowledge` and `stream` are **mesh-hosted
applications** — first-party, shipped with everything else, and running on the mesh the way
anything else does. A task does not need to know a node exists, and *being ours does not make
something infrastructure*. `ai` is folded into `config`: a provider licence is an ordinary grant.
**`record` is an open question rather than an eighth entry.** Contexts integrate through it
([ADR 0045](../../02-DECISIONS/0045-a-context-owns-its-store.md)), which makes it load-bearing,
and [research 006](../../01-RESEARCH/006-mesh-from-scratch/skeleton.md) leaves *where it lives*
unresolved — putting it in the substrate risks recreating the circularity the tier design just
removed. Listing it here would settle by naming what has not been settled by arguing.
**These are contexts, not services.** They are separate in the sense that matters — each owns
its own store, and they integrate through the record rather than by reading one another
@@ -118,9 +128,10 @@ every public name.
## Open
- **The contexts themselves.** Ten is the skeleton's claim, not a settled list. Research 006
asks whether the record belongs here or in the substrate, and whether identity is a context or
a substrate service.
- ~~**The contexts themselves.**~~ **Decided** — seven, by
[ADR 0055](../../02-DECISIONS/0055-the-control-plane-is-the-node-coordinating-contexts.md).
What remains open is narrower and named there: **where the record lives**, which research 006
leaves unresolved because the substrate is the one place it must not go.
- **How far it may be split.** One deployable today. Splitting a context out costs the single
interface a surface depends on
([research 011](../../01-RESEARCH/011-the-module-graph/00-overview.md)).
+6 -5
View File
@@ -17,14 +17,15 @@ document is written and this one's status becomes `implemented`.
| [`05-the-node-host.md`](05-the-node-host.md) | Tier 0 — the one thing installed by hand, and the only thing that changes a machine | [ADR 0037](../../02-DECISIONS/0037-the-host-applies-it-does-not-decide.md) |
| [`06-the-control-plane.md`](06-the-control-plane.md) | Tier 2 — what the term means, and the test for what belongs in it | [ADR 0037](../../02-DECISIONS/0037-the-host-applies-it-does-not-decide.md) |
| [`07-the-substrate.md`](07-the-substrate.md) | Tier 1 — what the control plane consumes and cannot grant itself | [ADR 0038](../../02-DECISIONS/0038-a-node-joins-by-linking-first.md), [0048](../../02-DECISIONS/0048-the-substrate-is-named.md) |
| [`08-connectivity.md`](08-connectivity.md) | One context in full — overlay, resolution, exposure, filtering, certificates | [ADR 0049](../../02-DECISIONS/0049-a-route-is-a-grant.md), [0050](../../02-DECISIONS/0050-reachability-is-a-property-of-the-address.md), [0051](../../02-DECISIONS/0051-the-enrolment-token-carries-the-mesh.md) |
| [`08-connectivity.md`](08-connectivity.md) | One context in full — overlay, resolution, exposure, filtering, certificates | [ADR 0049](../../02-DECISIONS/0049-a-route-is-a-grant.md), [0050](../../02-DECISIONS/0050-reachability-is-a-property-of-the-address.md), [0051](../../02-DECISIONS/0051-the-enrolment-token-carries-the-mesh.md), [0055](../../02-DECISIONS/0055-the-control-plane-is-the-node-coordinating-contexts.md) |
## Not yet written
- **The remaining contexts.** [ADR 0015](../../02-DECISIONS/0015-mesh-brokers-nodes-host-agents-think.md)
decides the decomposition; `connectivity` is the first written in full
([`08`](08-connectivity.md)) and the others do not exist yet. The work breakdown says in what
order they are needed.
- **The remaining six contexts.**
[ADR 0055](../../02-DECISIONS/0055-the-control-plane-is-the-node-coordinating-contexts.md)
settles the list at seven; `connectivity` is the first written in full
([`08`](08-connectivity.md)) and the other six do not exist yet. The work breakdown says in
what order they are needed.
- ~~**Domain grouping outside the core.**~~ **Not needed.**
[ADR 0017](../../02-DECISIONS/0017-modules-outside-the-core-are-grouped-by-domain.md) is
superseded by [ADR 0044](../../02-DECISIONS/0044-a-module-declares-presence-instantiation-and-exclusion.md):