feat: add a live Postgres city to the front page - #34

ci: add SECURITY.md and CodeQL analysis workflow for GitHub

See merge request postgres-ai/docs!883

Green: Add preview environments for docs

See merge request postgres-ai/docs!880

feat(blog): NOT EXISTS beats EXISTS by 20x — index size and heap fetches

Closes #183

See merge request postgres-ai/docs!885

feat(blog): how moving one word can speed up a query 10-50x

Closes #185

See merge request postgres-ai/docs!889

fix(blog): add equivalence note — PK + FK make the two queries identical

See merge request postgres-ai/docs!891

ops: DCF-5 add CI peer review gate (SOC2)

See merge request postgres-ai/docs!886

fix: bump DBLab Engine version references from 4.0.3 to 4.0.4

Closes #186

See merge request postgres-ai/docs!892

Blog: PG18 preserves planner stats during upgrades — even from PG14

Closes #187

See merge request postgres-ai/docs!893

fix: reduce blog feed size — summaries only, limit 20 entries

Closes #190

See merge request postgres-ai/docs!894

docs: Teleport integration howto for DBLab Engine

Closes database-lab#683

See merge request postgres-ai/docs!890

Add DBLab Engine 4.1 documentation: monitoring, RDS refresh, Teleport

See merge request postgres-ai/docs!896

Add DBLab Engine 4.1.0 release blog post

See merge request postgres-ai/docs!895

Fix xmin horizon data vs catalog query

Closes #192

See merge request postgres-ai/docs!898

fix: bump DBLab docs references to 4.1.1

See merge request postgres-ai/docs!899

ops: sunset old v2 staging — gate CI deploy jobs, REPLICAS=0

See merge request postgres-ai/docs!907

docs(supabase): embed PostgresAI for Supabase video on /supabase landing

Closes #234

See merge request postgres-ai/docs!947

docs: deep copy polish of full documentation tree

Closes #236

See merge request postgres-ai/docs!949

Remove careers pages

Closes #238

See merge request postgres-ai/docs!951

…master'

docs: drop DOCKER_API_VERSION pins, warn against publishing port 2346

Closes #239

See merge request postgres-ai/docs!952

Remove careers section

Closes #237

See merge request postgres-ai/docs!950

docs: pgai joe CLI (Joe API v2)

Closes #251

See merge request postgres-ai/docs!966

Remove Starter from public pricing and docs

Closes #252

See merge request postgres-ai/docs!968

docs: document PostgreSQL 19 beta support

Closes #253

See merge request postgres-ai/docs!969

docs(cli): document top-level login command

Closes #249

See merge request postgres-ai/docs!965

Text polish: docs site UI components (visible text)

Closes #248

See merge request postgres-ai/docs!964

…ration boundary

docs(monitoring): AAS health-matrix 1-minute peak granularity + calibration boundary

Closes #233

See merge request postgres-ai/docs!946

Text polish: landing page UI text (homepage, pricing, products)

Closes #247

See merge request postgres-ai/docs!963

Text polish: DBLab how-to guides (grammar, clarity, SEO)

Closes #242

See merge request postgres-ai/docs!958

An animated, simplified PostgreSQL cluster, drawn in canvas 2D from a small
deterministic model that runs in the reader's browser. No new dependency, no
WebGL, no network: 10 KiB gzipped for the scene and the model together.

Derived from PGSimCity (NikolayS/PGSimCity, Apache-2.0), which is the same
cluster in three dimensions. District positions, the compass and both semantic
palettes are that project's, so a reader who follows the link finds the same
colours meaning the same things.

Everything that moves is a consequence of model state, and the order it
happens in is the order PostgreSQL does the work:

  - The postmaster forks one backend per connection, not per statement. Each
    connection gets its own duct and keeps it for the session.
  - A read that misses in the pool goes to storage; the clock sweep picks the
    victim, caps usage_count at 5, and a dirty victim is written out first.
  - A write dirties a page in shared memory and produces a WAL record. The
    commit waits for that record to reach durable storage — and the page is
    still dirty in memory when the commit returns. No path marks a frame clean
    at commit time.
  - Dirty pages reach the data directory at a checkpoint, through the
    background writer, or because another backend needed the frame. A
    write-back only cleans a frame if it was not re-dirtied mid-write.
  - full_page_writes re-arms at every checkpoint, so WAL volume climbs after
    one begins, and replication lag climbs with it.
  - Replication ships the log. No data page crosses the wire.

The scene ships two hand-tuned palettes rather than one dimmed set, follows
data-theme, pauses off-screen and in a hidden tab, honours
prefers-reduced-motion with an explicit control to start it, and exposes every
district to the keyboard and to screen readers with a full text alternative.
Dirty frames also stand proud of the deck: clean and dirty are 140 degrees
apart in hue but close in luminance, so height carries the same information
where colour cannot.

Only dimensionless figures are shown — cache hit ratio by PostgreSQL's own
blks_hit / (blks_hit + blks_read), dirty share of the pool, and a replay-lag
bar — so nothing can be mistaken for measured throughput. The caption stating
that this is a model and not live data is content, not chrome, and is present
at every breakpoint.

`bun test src/components` asserts what the drawing claims: that no structure
stands unsupported over the excavation, that storage renders below memory,
that each backend has its own connection duct, that pages survive a commit
dirty, and that the palettes stay perceptually separable in both themes.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01NeTbYVtLkNTuwrRC6yUVva

fix(consulting): center CTA label

Closes #255

See merge request postgres-ai/docs!971

The branch was cut from the GitHub mirror's master, which last moved in
February — 120 commits and 309 files behind GitLab's master. That stale base
is why no preview environment appeared on !972: the branch carried February's
144-line .gitlab-ci.yml, which has no deploy_review job and no
merge_request_event rules at all, so the only pipeline it could produce was
the old push-triggered staging track. GitLab's master has carried the Green
per-MR preview environments for months.

Merging master in rather than rebasing, per the git-commit-standards rule
against rewriting pushed history.

No conflicts. The component and its mount survive unchanged, and the mount
still sits between the hero CTA and the demo video on the current, copy-edited
front page. bun test src/components: 47 pass. bun run build: succeeds.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01NeTbYVtLkNTuwrRC6yUVva