prole/docs/db-access.md
chrisfu 00f0ebec07 Merge claude/crazy-bose-fec256 into main
Bringing the long-running session-feature branch back into main in one
deliberate sweep. The branch carried the cluster work that's been live for
weeks (cross-cluster CNPG metrics, Grafana w/ Google OAuth, supabase
oauth2-proxy, cluster recovery, pg.0.knoe.dev + per-engineer onboarding,
GCS-backed CNPG backups via Workload Identity, the env-contamination
guard, the Junie brief queue, the cnpg-grafana CSRF + memory-request
fixes from today), while main accumulated Junie's parallel knoe-auth
Phase 2 OIDC work (full provider surface: discovery, authorize, token,
userinfo, JWKS, RS256 signing, code exchange, session services).

Key decision: the two branches did COMPETING rebrands off the same
starting point (5ba9b63, 2026-04-27):

  - claude branch (commit b355855, earlier): org.prole.authority.* →
                                              dev.knoe.auth.*
                                              (artifact renamed to
                                              knoe-auth.jar)
  - main (commit 9daa94b, recent): org.prole.authority.* →
                                    dev.knoe.authority.*
                                    (kept "authority" artifact name)

dev.knoe.auth wins: cluster runs from this name, the Maven artifact is
already knoe-auth.jar, and the broader rename is the documented
namespace direction (per ~/.claude/projects/-Users-chrisfu-dev-knoe-db/
memory/MEMORY.md). All of main's recent Phase 2 OIDC content was ported
from authority/src/.../dev/knoe/authority/ into
authority/src/.../dev/knoe/auth/ with package declarations rewritten.

== File-level resolution summary ==

Textual conflicts (4):

  authority/pom.xml
    - Took our artifactId="auth"
    - Took our branch's removal of spring-security-kerberos-client
      (verified: Junie's Phase 2 OIDC code does not import it; the dep
      was already-dead config)

  docs/pipeline-phases.md
    - Took our branch's "Phase 1 not started" status. Main had a
      misplaced " Complete" with a knoe-auth-Phase-1 commit ref
      in the autobuild Phase 1 section — different domain.

  docs/plans/knoe-auth-round-1.md
    - Took our branch's dev.knoe.auth file table (vs main's
      dev.knoe.authority listing). Pure rename mismatch.

  supabase/helm/knoe-supabase/templates/kong/config.yaml
    - Took our branch's onboard route + plain dashboard wiring.
      Main had an oauth2proxy.enabled toggle that put oauth2-proxy as
      a Kong upstream — but the deployed architecture (commit 25f1b2e)
      has oauth2-proxy in FRONT of Kong, not behind. Main's wrapper
      reflected an architecture that was never deployed.
    - Took our branch's removal of basic-auth from dashboard route
      (queue #15 brief still tracks the matching values.yaml /
      kong/deployment.yaml cleanup).

Java tree reconciliation (44 file-pairs):

  20 dual-path source files + 2 dual-path tests
    Body-identical between main's authority/ and our branch's auth/
    after stripping package decls — main's commit 9daa94b was a pure
    rebrand. Took our branch's auth/ version for all 22.

  8 main-only source files (Phase 2 OIDC), ported into auth/:
    web/JwksController.java
    web/OidcAuthorizeController.java
    web/OidcDiscoveryController.java
    web/OidcTokenController.java
    web/OidcUserInfoController.java
    session/OidcCodeService.java
    session/OidcTokenService.java
    session/SessionService.java

  12 main-only test files, ported into auth/:
    HealthControllerTest.java
    enroll/EnrollValueTypesTest.java
    enroll/EnrollmentControllerTest.java
    enroll/TotpServiceTest.java
    kerberos/KadminClientTest.java
    kerberos/KerberosSpnegoResultTest.java
    web/LoginControllerTest.java
    admin/AdminControllerTest.java
    user/PrincipalNormalizerTest.java
    regression/IdentityRegressionTest.java
    session/OidcCodeServiceTest.java
    session/SessionServiceTest.java

  Port mechanics: read main:authority/...<file> via git show, then sed
  rewrite `package dev.knoe.authority` → `package dev.knoe.auth` and
  `import dev.knoe.authority` → `import dev.knoe.auth`. Body content
  unchanged.

  authority/src/main/java/dev/knoe/authority/ — DELETED (duplicate)
  authority/src/test/java/dev/knoe/authority/  — DELETED (duplicate)

== Verification ==

- grep -rln '<<<<<<<' across .java/.md/.yaml/.yml/.sh/.xml/.tpl: clean
- find authority/src -path '*/dev/knoe/authority*': empty (subtree gone)
- grep 'package dev.knoe.authority' across repo: clean
- bash -n install.sh deploy.sh etc/preflight_kubecontext.sh: clean
- git ls-files -u | wc -l: 0 unmerged paths
- helm lint supabase/helm/knoe-supabase: pre-existing failure on
  studioIngress.enabled undefined in values.yaml (introduced by Junie
  on main; unrelated to this merge — flagging as follow-up).

== Followups (carried into TODO ranked queue or noted here) ==

  - helm lint failure: studioIngress block in values.yaml is missing
    enable flag; templates/studio/{ingress,oauth2proxy-deployment,
    oauth2proxy-service}.yaml all reference studioIngress.enabled with
    no default. Pre-existing on main; not introduced by this merge.
  - The five Junie briefs filed on this branch are now reachable from
    main at docs/plans/junie/{02,06,07,13,15}-*.md. Junie can pick them
    up in any order.
  - knoe-auth Phase 2 OIDC source (now at dev.knoe.auth.*) is not yet
    deployed to the cluster. Deployment is its own task.
  - The branch claude/crazy-bose-fec256 stays in place (worktree at
    .claude/worktrees/crazy-bose-fec256 may have ongoing context for
    Claude Code sessions). Safe to delete once next session starts
    cleanly from main.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-01 16:39:10 -07:00

7.0 KiB

Direct database access — psql + Rust apps

Per-engineer access to the knoe-db Postgres cluster (PG18 / Percona) at pg.0.knoe.dev:5432.

Phase 1 (current): per-engineer postgres roles with strong passwords (30-day rotation), TLS required, named-roles-only ACL on the public LB. Phase 2 (queued for Junie): replaces the password mechanism with PG18 native OAUTHBEARER (libpq drives a Google Device Flow) + a small validator library bundled into our knoe-db image. Connection target stays the same; engineers' psql invocation only changes from -W to setting oauth_issuer / oauth_client_id in the connection string.


First-day flow (engineer-facing)

See docs/onboarding.md — the canonical engineer checklist. Two-line summary:

  1. chrisfu runs etc/onboard_engineer.sh <username> <email>. Script outputs a one-time URL (and QR-code rendering of it) plus a 24h temp password.
  2. Engineer opens URL on phone or laptop → page at db.0.knoe.dev/onboard.html shows their password + connection string with [Copy] buttons. Engineer pastes into psql, immediately rotates via \password, saves their new password to personal 1Password (no shared vault required).

The CA cert lives at etc/knoe-db-ca.crt in this repo — engineers reference it with sslrootcert=etc/knoe-db-ca.crt after cloning. It's a public CA root, not a secret. Auto-rotates with CNPG; we re-export and commit when it does.

Connection string (engineers substitute their own role):

psql "host=pg.0.knoe.dev port=5432 user=$USER dbname=postgres \
      sslmode=verify-full sslrootcert=etc/knoe-db-ca.crt"

Rust apps (sqlx / tokio-postgres / deadpool-postgres)

Connection string with the same TLS configuration:

let url = format!(
    "postgres://{user}:{password}@pg.0.knoe.dev:5432/postgres\
     ?sslmode=verify-full&sslrootcert={ca_path}",
    user = std::env::var("KNOE_PG_USER")?,
    password = std::env::var("KNOE_PG_PASSWORD")?,
    ca_path = format!("{}/.knoe/knoe-db-ca.crt", std::env::var("HOME")?),
);

let pool = sqlx::PgPool::connect(&url).await?;

For tokio-postgres directly, configure the MakeTlsConnector with the same CA bundle (use postgres-native-tls or tokio-postgres-rustls).

Service apps on GKE (Phase 2.5 territory): the same connection string, but KNOE_PG_USER / KNOE_PG_PASSWORD come from a per-app k8s Secret rather than 1Password. Phase 2 swaps this to Workload-Identity-issued OAUTHBEARER tokens — the app code change is one line.


What can a knoe_developer do?

The per-engineer roles (chrisfu, ron) are members of the knoe_developer group. Grants:

Schema Permission Notes
knoe SELECT, INSERT, UPDATE, DELETE Application schema; ALTER DEFAULT PRIVILEGES propagates to new tables
public SELECT, INSERT, UPDATE, DELETE Same
auth SELECT only Read-only on Supabase GoTrue tables (debugging / inspection)
storage SELECT only Read-only on Supabase storage tables
extensions USAGE Lets queries reach extensions.pg_stat_statements etc.

What you can NOT do:

  • Drop / truncate auth.* or storage.* (read-only)
  • ALTER ROLE, CREATE ROLE, CREATE EXTENSION (those are postgres-only)
  • Connect as postgres, supabase_admin, authenticator from outside the cluster
  • Connect at all without TLS (pg_hba hostnossl ... reject)

Need broader access? Ask chrisfu — easier to extend the knoe_developer group than to mint role-specific grants per engineer.


Onboarding a new engineer (admin-facing)

./etc/onboard_engineer.sh <username> <email>

Outputs:

  • A one-time URL: https://db.0.knoe.dev/onboard.html#user=...&pw=...&exp=...
  • A QR-code rendering of that URL (in your terminal, via qrencode)
  • The plaintext temp password (fallback)

Deliver the URL to the new engineer via Gmail (to their @knoey.com address) or by showing the QR on screenshare for them to scan with their phone. The landing page handles all the engineer-side UX (copy buttons, connection string, rotation guidance). They save the password to their personal 1Password — no shared vault required. See docs/onboarding.md.

The script idempotently CREATEs / ALTERs the role:

  • LOGIN INHERIT VALID UNTIL <now+24h>
  • PASSWORD set to a fresh 192-bit random
  • GRANT knoe_developer

The pg_hba.conf hostssl all +knoe_developer all scram-sha-256 rule means the new role is reachable via pg.0.knoe.dev immediately — no pg_hba edits needed per onboard.

Other actions:

./etc/onboard_engineer.sh --rotate <username>     # fresh 24h temp password
./etc/onboard_engineer.sh --revoke <username>     # DROP ROLE

Rotate your password

The engineer-side path is \password inside psql — they pick a new password locally, postgres updates the SCRAM verifier, the password never leaves their machine. No coordination with chrisfu needed.

If chrisfu needs to issue a fresh temp password (engineer locked out, forgot to rotate before VALID UNTIL expired):

./etc/onboard_engineer.sh --rotate <username>

Same delivery (URL/QR/email) as initial onboarding.


Troubleshooting

  • pg_hba.conf rejects connection for host ... no encryption — you set sslmode=disable. Use verify-full (requires the CA cert) or require (encrypted but not verified).

  • pg_hba.conf rejects connection for host <ip>, user <u>, no encryption — you're on a network that's stripping TLS, OR your psql client is too old. Update libpq to ≥17.

  • password authentication failed for user "<u>" — typically wrong password, but also fires for roles not in our allow list (postgres, knoe-db, etc.) when connecting from outside the cluster. Check you're using your own role.

  • FATAL: role "<u>" is not permitted to log in — your VALID UNTIL lapsed. Ask for rotation.


  • Phase 2 plan (queued, owned by Junie): install/deploy.sh switches to pg_oauth as the default until knoey-system/authority auth is live. Includes building a validator library against PG18's OAuthValidatorCallbacks API and bundling into the knoe-db image. See docs/TODO.md follow-up for the active reference.
  • k3d build: same Service + cluster cert pattern in k3d mode. Owned by Junie alongside the install.sh true-up.
  • knoe-auth Round 1: the longer-term identity story (Kerberos KDC + Java authority). Becomes the Phase 2 issuer when it deploys, but Phase 2 doesn't block on it — Google's issuer is sufficient for now.

Files