mirror of
https://scm.tikali.ai/tikali/applications/monky/monky-deployd.git
synced 2026-09-18 05:36:15 +00:00
feat: monky-deployd v0.1.0 — pull agent over the mesh (ADR-0028)
Stdlib-only Python 3.12 agent for docker VMs and laptops: flock → checkin
(bearer = the agent's OpenBao token, bootstrapped from the install kit's
jwt-tenancy deploy grant) → action apply|none|down → bundle (sha256
verified) → refusal checks (unresolved ${VAR} names only, manifest paths
pinned to monky/data/<env>/see/, privileged/host-network, rollback, disk
need×1.5+headroom) → lease → POST /v1/auth/jwt-tenancy/login → KV reads →
.env 0600 → promote → compose pull/up → wait healthy → report; finally
renew-self / re-lease before max TTL, scrub. Exit 0/75/78/1. Redactor log
filter. Transports sdk (openziti) / proxy (ziti tunnel proxy 18443/18200) /
system. Laptop mode.
Packaging: hardened oneshot + 60 s timer + proxy unit, nfpm .deb with
/opt/monky-deployd/venv, install.sh for Ubuntu 26.04 (Gitea release
download, enrol, ACLs, bootstrap from stdin), ansible role skeleton for
osg1-07. CI: lint/test on every change; wheel (openziti on ubuntu:26.04) and
package (nfpm) allow_failure until runner egress is proven; GitLab release +
release:gitea on v* tags. Docs: README, PROTOCOL, OPERATIONS, CHANGELOG,
CLAUDE/AGENTS.
Divergence noted: monky-tenancy main (MR !15) still ships the AppRole lease
and kit; this agent implements the plan's Gate 1 RESULT (login_jwt, no
unwrap) and refuses an AppRole lease loudly (LEASE_SHAPE).
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01KLB7jieMNRkTsJ2epr4Ds1
This commit is contained in:
@@ -0,0 +1,109 @@
|
||||
<!-- xlate:verbatim-fences -->
|
||||
# Operating monky-deployd
|
||||
|
||||
## Units
|
||||
|
||||
| unit | what |
|
||||
|---|---|
|
||||
| `monky-deployd.timer` | fires `monky-deployd.service` 60 s after the previous tick finished (+ ≤10 s jitter); `OnBootSec=90s` |
|
||||
| `monky-deployd.service` | `Type=oneshot`, `monky-deployd run --once` as user `monky-deployd` (+ `docker` group); `SuccessExitStatus=75`; hardened (`NoNewPrivileges`, `ProtectSystem=strict`, `ReadWritePaths=/var/lib/monky-deployd /etc/monky-deployd /run/docker.sock`, `UMask=0077`, no capabilities) |
|
||||
| `monky-deployd-proxy.service` | only with `transport: proxy`: `ziti tunnel proxy -i <identity> monky.tenancy.deploy:18443 openbao:18200` as user `ziti`; `EnvironmentFile=/etc/monky-deployd/proxy.env` |
|
||||
| `ziti-edge-tunnel.service` | the host identity's tunneler in **`run-host`** mode (drop-in `run-host.conf` written by `install.sh`) |
|
||||
|
||||
```sh
|
||||
systemctl status monky-deployd.timer monky-deployd.service
|
||||
systemctl list-timers monky-deployd.timer
|
||||
systemctl start monky-deployd.service # tick now (blocks until done)
|
||||
journalctl -u monky-deployd -f # priorities are real (INFO/WARNING/ERROR)
|
||||
monky-deployd status # exit 0 = token present and in sync
|
||||
monky-deployd status --json | jq .
|
||||
```
|
||||
|
||||
Files: `/etc/monky-deployd/config.yaml` (0640 root:monky-deployd), `/etc/monky-deployd/openbao-ca.pem`,
|
||||
`/etc/monky-deployd/bootstrap.jwt` (only until the first login), `/var/lib/monky-deployd/{bao.token,state.json,lock}`,
|
||||
`/var/lib/monky-deployd/<env>/releases/<sha>/` + `current` (the compose project dir, `.env` 0600),
|
||||
`/opt/openziti/etc/identities/monky-host.<env>.json` (ziti:ziti 0600 + ACL `u:monky-deployd:r`).
|
||||
|
||||
## Reading the journal
|
||||
|
||||
| line | meaning |
|
||||
|---|---|
|
||||
| `checkin: action=none desired=… applied=…` then `healthy; heartbeat reported` | converged |
|
||||
| `checkin: action=apply …` → `read 3 secret(s): GEMINI_API_KEY, …` → `promoted release …` → `applied …` | a deploy |
|
||||
| `refused: ENV_INCOMPLETE: unresolved: X` | the bundle needs a variable no manifest entry supplies — fix the descriptor / set the secret in the console; nothing was started |
|
||||
| `refused: DISK_INSUFFICIENT: docker data-root has N MiB free, bundle needs M MiB` | free space (the env-dev-09 lesson): grow the data-root disk or prune |
|
||||
| `refused: PRIVILEGED_REFUSED` / `ROLLBACK_REFUSED` | the bundle needs `allow_privileged` / `allow_rollback` in its `agent` profile |
|
||||
| `temporary network failure` (exit 75) | mesh/tenancy unreachable — check `ziti-edge-tunnel`, the identity's terminators, `monky.tenancy.deploy` health |
|
||||
| `AGENT_UNAUTHENTICATED: bearer refused` (exit 1) | the grant was superseded (kit re-revealed / retire) or the token revoked → re-run the install kit |
|
||||
| `AGENT_ENV_MISMATCH` (exit 78) | the token belongs to another env than `config.yaml` — fix the config or re-issue the identity; the timer keeps firing but every tick exits 78 immediately (no storm) |
|
||||
| `failed: docker compose pull failed (rc=1)` | registry/pull problem; compose output is in the report's tail and in the journal |
|
||||
|
||||
Values never appear in the journal (`[REDACTED]` for token shapes and every value the agent has read).
|
||||
|
||||
## Secrets on the box
|
||||
|
||||
The only credential is `/var/lib/monky-deployd/bao.token` (0600, user `monky-deployd`). `.env`
|
||||
files under `releases/<sha>/` hold the rendered values (0600, same user, `docker compose` reads
|
||||
them). Rotating a secret in the console changes the bundle sha → the next tick re-reads and
|
||||
re-applies. To force a re-read now: `systemctl start monky-deployd.service`.
|
||||
|
||||
Lost or revoked token: `monky-deployd status` shows `token ABSENT`; reveal the kit again (admin,
|
||||
`GET /v1/backends/{id}/agent/install`), paste its bootstrap grant to `/etc/monky-deployd/bootstrap.jwt`
|
||||
(0600 monky-deployd) or re-run `install.sh` (enrolment is skipped when the identity exists), then
|
||||
`monky-deployd bootstrap` or wait a tick.
|
||||
|
||||
## Retire and volumes policy
|
||||
|
||||
Retire is driven by tenancy (`POST /v1/backends/{id}/retire {confirm, force, purge_volumes}`):
|
||||
the next check-in returns `action: down` and the agent runs `docker compose down --remove-orphans`.
|
||||
Data volumes:
|
||||
|
||||
- `-v` (remove volumes) only when tenancy sent `purge_volumes: true` **or** the box's
|
||||
`volumes_on_absent: purge` — and **never on a prod env** (the request is logged and ignored).
|
||||
- default `volumes_on_absent: keep`: volumes stay for a manual `docker volume rm` later.
|
||||
|
||||
After `down` the agent reports `down`, clears `applied_sha`, removes `current` and keeps
|
||||
`releases/` (no secrets outside `.env`, which you can `shred`). Tenancy then destroys the Bao
|
||||
paths, revokes the token accessor and deletes the host identity → the following ticks fail with
|
||||
`AGENT_UNAUTHENTICATED`; `systemctl disable --now monky-deployd.timer` and `apt remove monky-deployd`
|
||||
(`apt purge` also removes `/var/lib/monky-deployd` and `/etc/monky-deployd`).
|
||||
|
||||
## Laptop mode
|
||||
|
||||
`laptop_mode: true` (`install.sh --laptop`): being offline is normal — a tick that cannot reach the
|
||||
mesh logs one INFO line and exits 0. Run without the system timer:
|
||||
|
||||
```sh
|
||||
monky-deployd run # loop, interval_s + jitter, SIGTERM stops; or a user timer with `run --once`
|
||||
```
|
||||
|
||||
Liveness grace for laptops is 24 h on the tenancy side (P0); `offline` only hides the backend from
|
||||
the picker, nothing is retired automatically.
|
||||
|
||||
## Housekeeping
|
||||
|
||||
`monky-deployd run --once --prune` removes every `releases/<sha>` except `current` and runs
|
||||
`docker image prune -f` after a successful apply. The state dir and the token survive `apt remove`;
|
||||
`apt purge` deletes them.
|
||||
|
||||
## Publishing a release to Gitea by hand
|
||||
|
||||
When `release:gitea` is manual (no `GITEA_TOKEN` in CI) — from jump1 with the Gitea API token:
|
||||
|
||||
```sh
|
||||
T=$(tr -d '\n' < ~/.gitlab_tokens/gitea-token); G=http://172.16.8.1:3000/api/v1/repos/mdella/monky-deployd
|
||||
curl -s -X POST -H "Authorization: token $T" $G/mirror-sync # pull the tag
|
||||
RID=$(curl -s -X POST -H "Authorization: token $T" -H 'Content-Type: application/json' $G/releases \
|
||||
-d '{"tag_name":"v0.1.0","name":"monky-deployd v0.1.0","body":"See CHANGELOG.md"}' | jq -r .id)
|
||||
for f in monky-deployd_0.1.0_amd64.deb monky-deployd_0.1.0_amd64.deb.sha256 install.sh; do
|
||||
curl -s -X POST -H "Authorization: token $T" -F "attachment=@$f" "$G/releases/$RID/assets?name=$f"; done
|
||||
```
|
||||
|
||||
The assets are then at `https://gitea.cbs.tikali.net/mdella/monky-deployd/releases/download/v0.1.0/<file>`,
|
||||
which is what `install.sh` fetches (GitLab artifacts from the tag pipeline's `package` job).
|
||||
|
||||
## Ansible (osg1-07)
|
||||
|
||||
`ansible/roles/monky_deployd/` is the role skeleton to copy into osg1-07: `.deb` from the Gitea
|
||||
release (sha256-verified), config template, bootstrap grant from a vault var, ACL on the identity,
|
||||
timer, one tick. Rolled to env-dev-06..09 after the env-qa-02 pilot; env-dev-01 last.
|
||||
Reference in New Issue
Block a user