From d74ce32cd0ea68d86cce84a51b1d1ab1cf4c6f08 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ullrich=20Sch=C3=A4fer?= <hi+github@ullrich.is>
Date: Sun, 7 Jun 2026 08:43:03 +0200
Subject: [PATCH 1/3] fix(journal): actor profile fields must serialize as a
 JSON array
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Mastodon's PropertyValue parser requires the actor's attachment to be
a JSON *array* and silently ignores a bare object — and Fedify
compacts single-element arrays to bare objects, so the 🥾 trails.cool
field shipped in #464 never rendered (verified: Mastodon stored
fields = {} after a forced actor refresh on the soak instance).

Ship two fields so the array survives serialization: the 🥾 profile
link (rel=me) plus an Instance link — the latter is genuinely useful
for self-hosted instances anyway. Test now asserts the array shape
explicitly.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
---
 apps/journal/app/lib/federation.server.test.ts | 18 ++++++++++++------
 apps/journal/app/lib/federation.server.ts      |  8 ++++++++
 2 files changed, 20 insertions(+), 6 deletions(-)

diff --git a/apps/journal/app/lib/federation.server.test.ts b/apps/journal/app/lib/federation.server.test.ts
index f00aa39..294ab08 100644
--- a/apps/journal/app/lib/federation.server.test.ts
+++ b/apps/journal/app/lib/federation.server.test.ts
@@ -95,12 +95,18 @@ describe("actor object", () => {
     expect(actor.preferredUsername).toBe("bruno");
     expect(actor.name).toBe("Bruno");
     expect(actor.summary).toBe("Riding bikes");
-    // Profile-metadata field Mastodon renders ("this is a trails profile")
-    const attachments = Array.isArray(actor.attachment) ? actor.attachment : [actor.attachment];
-    const field = attachments.find((a: { type: string }) => a?.type === "PropertyValue");
-    expect(field?.name).toBe("🥾 trails.cool");
-    expect(field?.value).toContain('href="http://localhost:3000/users/bruno"');
-    expect(field?.value).toContain('rel="me"');
+    // Profile-metadata fields Mastodon renders ("this is a trails
+    // profile"). MUST serialize as a JSON array — Mastodon's parser
+    // silently ignores a bare attachment object, which is what Fedify
+    // compacts single-element arrays into (hence two fields).
+    expect(Array.isArray(actor.attachment)).toBe(true);
+    const fields = actor.attachment.filter((a: { type: string }) => a?.type === "PropertyValue");
+    expect(fields.length).toBeGreaterThanOrEqual(2);
+    const trails = fields.find((f: { name: string }) => f.name === "🥾 trails.cool");
+    expect(trails?.value).toContain('href="http://localhost:3000/users/bruno"');
+    expect(trails?.value).toContain('rel="me"');
+    const instance = fields.find((f: { name: string }) => f.name === "Instance");
+    expect(instance?.value).toContain("localhost:3000");
   });
 
   it("404s the actor for a private user", async () => {
diff --git a/apps/journal/app/lib/federation.server.ts b/apps/journal/app/lib/federation.server.ts
index 6cb971e..8622653 100644
--- a/apps/journal/app/lib/federation.server.ts
+++ b/apps/journal/app/lib/federation.server.ts
@@ -180,11 +180,19 @@ function buildFederation(): Federation<void> {
         // metadata table — a human-visible "this is a trails profile"
         // marker. (The machine-readable marker is NodeInfo; this is
         // flair.) Mastodon strips most HTML in values but keeps links.
+        // NOTE: Mastodon's parser requires `attachment` to be a JSON
+        // *array* and silently ignores a bare object — and Fedify
+        // compacts single-element arrays to bare objects. Keep at least
+        // two fields here so the array survives serialization.
         attachments: [
           new PropertyValue({
             name: "🥾 trails.cool",
             value: `<a href="${localActorIri(identifier)}" rel="me">${getOrigin().replace(/^https?:\/\//, "")}/users/${identifier}</a>`,
           }),
+          new PropertyValue({
+            name: "Instance",
+            value: `<a href="${getOrigin()}">${getOrigin().replace(/^https?:\/\//, "")}</a>`,
+          }),
         ],
       });
     })

From 4e63f7631a75d3ccf8808c803b73a9bb1a7d3ead Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ullrich=20Sch=C3=A4fer?= <hi+github@ullrich.is>
Date: Sun, 7 Jun 2026 09:13:44 +0200
Subject: [PATCH 2/3] fix(journal): show remote followers in follower/following
 lists
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

countFollowers counts every accepted follow row, but listFollowers
inner-joined users on follower_id — so a federated follower (NULL
follower_id, follower_actor_iri set) bumped the count while never
appearing in the list. Observed live: profile said '1 follower', list
below was empty. listFollowing had the same latent bug for outbound
trails-to-trails follows.

Both lists now left-join users + remote_actors: local entries link to
the local profile as before; remote entries display cached actor data
(IRI parsing as fallback) and link out to the remote profile.

Integration test pins count/list consistency for a remote follower.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
---
 .../journal/app/components/CollectionPage.tsx |  8 +-
 .../lib/federation-inbox.integration.test.ts  | 17 ++++
 apps/journal/app/lib/follow.server.ts         | 99 ++++++++++++++++---
 3 files changed, 109 insertions(+), 15 deletions(-)

diff --git a/apps/journal/app/components/CollectionPage.tsx b/apps/journal/app/components/CollectionPage.tsx
index 08c8bab..2632b44 100644
--- a/apps/journal/app/components/CollectionPage.tsx
+++ b/apps/journal/app/components/CollectionPage.tsx
@@ -4,6 +4,9 @@ interface Entry {
   username: string;
   displayName: string | null;
   domain: string;
+  /** Local path (`/users/x`) or, for federated entries, the remote profile URL. */
+  profileUrl: string;
+  remote: boolean;
 }
 
 interface Props {
@@ -42,9 +45,10 @@ export function CollectionPage({ kind, user, entries, page, total }: Props) {
       ) : (
         <ul className="mt-6 divide-y divide-gray-200 rounded-lg border border-gray-200 bg-white">
           {entries.map((entry) => (
-            <li key={entry.username} className="px-4 py-3">
+            <li key={`${entry.username}@${entry.domain}`} className="px-4 py-3">
               <a
-                href={`/users/${entry.username}`}
+                href={entry.profileUrl}
+                {...(entry.remote ? { target: "_blank", rel: "noopener noreferrer" } : {})}
                 className="flex items-center justify-between hover:underline"
               >
                 <span className="text-sm font-medium text-gray-900">
diff --git a/apps/journal/app/lib/federation-inbox.integration.test.ts b/apps/journal/app/lib/federation-inbox.integration.test.ts
index 313c0b4..9adcbee 100644
--- a/apps/journal/app/lib/federation-inbox.integration.test.ts
+++ b/apps/journal/app/lib/federation-inbox.integration.test.ts
@@ -64,6 +64,23 @@ describe.runIf(runIntegration)("federation inbox (integration)", () => {
     expect(rows[0]!.acceptedAt).not.toBeNull();
   });
 
+  it("remote followers appear in the followers list (count/list consistency)", async () => {
+    const username = `fed-list-${Date.now()}`;
+    const userId = await makeUser({ username });
+    await recordRemoteFollow(REMOTE_ACTOR, username);
+
+    const { listFollowers, countFollowers } = await import("./follow.server.ts");
+    const [entries, total] = await Promise.all([listFollowers(userId), countFollowers(userId)]);
+    expect(total).toBe(1);
+    expect(entries).toHaveLength(1);
+    const entry = entries[0]!;
+    expect(entry.remote).toBe(true);
+    expect(entry.profileUrl).toBe(REMOTE_ACTOR);
+    // No remote_actors cache row in this test → identity parsed from the IRI.
+    expect(entry.username).toBe("alice");
+    expect(entry.domain).toBe("other-trails.example");
+  });
+
   it("inbound Follow is idempotent — replays don't double-insert", async () => {
     const username = `fed-replay-${Date.now()}`;
     const userId = await makeUser({ username });
diff --git a/apps/journal/app/lib/follow.server.ts b/apps/journal/app/lib/follow.server.ts
index f1254ae..0105e5d 100644
--- a/apps/journal/app/lib/follow.server.ts
+++ b/apps/journal/app/lib/follow.server.ts
@@ -1,7 +1,7 @@
 import { randomUUID } from "node:crypto";
 import { eq, and, count, desc, isNull, isNotNull } from "drizzle-orm";
 import { getDb } from "./db.ts";
-import { users, follows } from "@trails-cool/db/schema/journal";
+import { users, follows, remoteActors } from "@trails-cool/db/schema/journal";
 import { localActorIri } from "./actor-iri.ts";
 import { createNotification } from "./notifications.server.ts";
 import { logger } from "./logger.server.ts";
@@ -283,35 +283,89 @@ export interface CollectionEntry {
   username: string;
   displayName: string | null;
   domain: string;
+  /** Where the entry's profile lives: local path or remote actor URL. */
+  profileUrl: string;
+  remote: boolean;
 }
 
 const COLLECTION_PAGE_SIZE = 50;
 
+/**
+ * Best-effort identity for a remote actor we may or may not have
+ * cached: prefer the remote_actors row, fall back to parsing the IRI
+ * (canonical fediverse shape `https://host/users/name`).
+ */
+function remoteEntry(
+  actorIri: string,
+  cached: { username: string | null; displayName: string | null; domain: string | null },
+): CollectionEntry {
+  let host = "";
+  let lastSegment: string;
+  try {
+    const url = new URL(actorIri);
+    host = url.host;
+    lastSegment = url.pathname.split("/").filter(Boolean).pop() ?? "";
+  } catch {
+    // Malformed IRI in the DB — display it raw rather than hide the row.
+    lastSegment = actorIri;
+  }
+  return {
+    username: cached.username ?? lastSegment,
+    displayName: cached.displayName,
+    domain: cached.domain ?? host,
+    profileUrl: actorIri,
+    remote: true,
+  };
+}
+
 /**
  * Paginated list of accepted followers of `userId`. Newest acceptance first.
- * Pending requests are excluded — they live in the Requests tab on
- * /notifications.
+ * Includes remote (federated) followers — `follower_actor_iri` rows — so the
+ * list matches countFollowers; display data comes from the remote_actors
+ * cache when present, IRI parsing otherwise. Pending requests are excluded —
+ * they live in the Requests tab on /notifications.
  */
 export async function listFollowers(userId: string, page: number = 1): Promise<CollectionEntry[]> {
   const db = getDb();
   const offset = (Math.max(1, page) - 1) * COLLECTION_PAGE_SIZE;
   const rows = await db
     .select({
-      username: users.username,
-      displayName: users.displayName,
-      domain: users.domain,
+      localUsername: users.username,
+      localDisplayName: users.displayName,
+      localDomain: users.domain,
+      followerActorIri: follows.followerActorIri,
+      remoteUsername: remoteActors.username,
+      remoteDisplayName: remoteActors.displayName,
+      remoteDomain: remoteActors.domain,
     })
     .from(follows)
-    .innerJoin(users, eq(follows.followerId, users.id))
+    .leftJoin(users, eq(follows.followerId, users.id))
+    .leftJoin(remoteActors, eq(follows.followerActorIri, remoteActors.actorIri))
     .where(and(eq(follows.followedUserId, userId), isNotNull(follows.acceptedAt)))
     .orderBy(desc(follows.acceptedAt))
     .limit(COLLECTION_PAGE_SIZE)
     .offset(offset);
-  return rows;
+  return rows.map((r) =>
+    r.localUsername !== null
+      ? {
+          username: r.localUsername,
+          displayName: r.localDisplayName,
+          domain: r.localDomain ?? "",
+          profileUrl: `/users/${r.localUsername}`,
+          remote: false,
+        }
+      : remoteEntry(r.followerActorIri ?? "", {
+          username: r.remoteUsername,
+          displayName: r.remoteDisplayName,
+          domain: r.remoteDomain,
+        }),
+  );
 }
 
 /**
  * Paginated list of accepted follows from `userId`. Newest acceptance first.
+ * Includes remote (trails-to-trails) targets — rows whose followed side is
+ * only an actor IRI — for the same count/list consistency as listFollowers.
  * Pending outgoing follows (against private/locked targets) are excluded.
  */
 export async function listFollowing(userId: string, page: number = 1): Promise<CollectionEntry[]> {
@@ -319,15 +373,34 @@ export async function listFollowing(userId: string, page: number = 1): Promise<C
   const offset = (Math.max(1, page) - 1) * COLLECTION_PAGE_SIZE;
   const rows = await db
     .select({
-      username: users.username,
-      displayName: users.displayName,
-      domain: users.domain,
+      localUsername: users.username,
+      localDisplayName: users.displayName,
+      localDomain: users.domain,
+      followedActorIri: follows.followedActorIri,
+      remoteUsername: remoteActors.username,
+      remoteDisplayName: remoteActors.displayName,
+      remoteDomain: remoteActors.domain,
     })
     .from(follows)
-    .innerJoin(users, eq(follows.followedUserId, users.id))
+    .leftJoin(users, eq(follows.followedUserId, users.id))
+    .leftJoin(remoteActors, eq(follows.followedActorIri, remoteActors.actorIri))
     .where(and(eq(follows.followerId, userId), isNotNull(follows.acceptedAt)))
     .orderBy(desc(follows.acceptedAt))
     .limit(COLLECTION_PAGE_SIZE)
     .offset(offset);
-  return rows;
+  return rows.map((r) =>
+    r.localUsername !== null
+      ? {
+          username: r.localUsername,
+          displayName: r.localDisplayName,
+          domain: r.localDomain ?? "",
+          profileUrl: `/users/${r.localUsername}`,
+          remote: false,
+        }
+      : remoteEntry(r.followedActorIri, {
+          username: r.remoteUsername,
+          displayName: r.remoteDisplayName,
+          domain: r.remoteDomain,
+        }),
+  );
 }

From 89caedca058a7bb17901b2cb8e1986fa5de30eb9 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ullrich=20Sch=C3=A4fer?= <hi+github@ullrich.is>
Date: Sun, 7 Jun 2026 09:52:49 +0200
Subject: [PATCH 3/3] ci: scheduled image prune + staging-deploy disk hygiene
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

The 2026-06-07 disk-full outage (flagship 100%, postgres crash-looping
on its pidfile, prod + staging down): cd-apps DOES prune images after
deploys, but it had been failing early all day (migration bug), so its
prune never ran — while ~10 staging/preview deploys kept pulling fresh
images with no prune of their own.

- cd-staging: prune superseded layers (until=1h guard against racing
  in-flight pulls) after persistent staging and preview deploys.
- disk-maintenance.yml: NEW daily scheduled prune (04:30 UTC) that
  also FAILS when the disk is still ≥85% after pruning — a redundant
  alert channel for exactly the case where the Grafana disk alert
  drowns in other noise, as it did during the incident.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
---
 .github/workflows/cd-staging.yml       | 11 ++++++
 .github/workflows/disk-maintenance.yml | 52 ++++++++++++++++++++++++++
 2 files changed, 63 insertions(+)
 create mode 100644 .github/workflows/disk-maintenance.yml

diff --git a/.github/workflows/cd-staging.yml b/.github/workflows/cd-staging.yml
index 7b87971..ed59b70 100644
--- a/.github/workflows/cd-staging.yml
+++ b/.github/workflows/cd-staging.yml
@@ -195,6 +195,13 @@ jobs:
             # via cd-infra) are live. Idempotent.
             docker compose exec -T caddy caddy reload --config /etc/caddy/Caddyfile || true
 
+            # Reclaim superseded image layers. cd-apps prunes after its own
+            # deploys, but a day of staging/preview deploys while cd-apps is
+            # red can fill the disk on its own (2026-06-07: 100% full,
+            # postgres down). The 1h filter avoids racing layers another
+            # in-flight deploy just pulled.
+            docker image prune -af --filter "until=1h" || true
+
             docker compose -f docker-compose.staging.yml -p trails-staging --env-file staging.env --profile persistent ps
 
   # ── PR preview deploy ────────────────────────────────────────────────
@@ -356,6 +363,10 @@ jobs:
             # Reload Caddy to pick up the per-PR snippet (writes/replaces it from the SCP step)
             docker compose exec -T caddy caddy reload --config /etc/caddy/Caddyfile
 
+            # Same disk-hygiene prune as the persistent staging deploy —
+            # preview pushes are the highest-volume image source.
+            docker image prune -af --filter "until=1h" || true
+
             docker compose -f docker-compose.staging.yml -p "$PROJECT" --env-file "$ENV_FILE" ps
 
       # Find any prior preview comment so we can update it in place rather
diff --git a/.github/workflows/disk-maintenance.yml b/.github/workflows/disk-maintenance.yml
new file mode 100644
index 0000000..5113a0c
--- /dev/null
+++ b/.github/workflows/disk-maintenance.yml
@@ -0,0 +1,52 @@
+name: Disk Maintenance
+
+# Daily safety net for flagship disk usage. The deploy workflows prune
+# superseded image layers after their own runs, but that protection
+# disappears exactly when it's needed most: a deploy that fails early
+# never reaches its prune step, while other workflows keep pulling
+# fresh images (2026-06-07 incident: cd-apps red all day, ~10 staging
+# deploys, disk 100% full, postgres down on a Saturday morning).
+#
+# Also doubles as a redundant alert channel: the run FAILS when the
+# disk is still above the threshold after pruning, so a scheduled-run
+# failure email lands even if the Grafana disk alert drowns in other
+# noise (which is what happened during the incident).
+
+on:
+  schedule:
+    # Daily at 04:30 UTC (offset from staging-cleanup's Monday 04:00)
+    - cron: "30 4 * * *"
+  workflow_dispatch: {}
+
+concurrency:
+  group: disk-maintenance
+  cancel-in-progress: false
+
+jobs:
+  prune-flagship:
+    name: Prune unused images (flagship)
+    runs-on: ubuntu-latest
+    environment: production
+    steps:
+      - name: Prune and check disk headroom
+        uses: appleboy/ssh-action@v1
+        with:
+          host: ${{ secrets.DEPLOY_HOST }}
+          username: root
+          key: ${{ secrets.DEPLOY_SSH_KEY }}
+          script: |
+            set -euo pipefail
+            echo "before: $(df -h / | tail -1)"
+            # 12h filter: never touch layers a same-day deploy may still
+            # be assembling; running containers' images are never pruned.
+            docker image prune -af --filter "until=12h"
+            echo "after:  $(df -h / | tail -1)"
+
+            PCT=$(df --output=pcent / | tail -1 | tr -dc '0-9')
+            if [ "$PCT" -ge 85 ]; then
+              echo "Disk still at ${PCT}% after pruning — needs a human."
+              echo "Largest docker consumers:"
+              docker system df
+              exit 1
+            fi
+            echo "Disk at ${PCT}% — healthy."