docs(adr): fix ADR-028 §12 false API-key claim, stale TBD, and §7 title

§12 stated OLLAMA_API_KEY guards against lateral movement — contradicts
§7's empirical finding that it is not enforced. Replaced with an accurate
note referencing §7. Stale pre-merge placeholder in Consequences ("Three
TBD items must be resolved") removed; all three are resolved. §7 section
title updated from "0.6.5" to "0.6.5 and 0.30.6" to match the body text.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
This commit is contained in:
Marcel
2026-06-06 14:37:03 +02:00
parent 5a939d9222
commit 3536ed884c

View File

@@ -110,7 +110,7 @@ if (!apiKey.isBlank()) {
Sending `Authorization: Bearer ` (empty token) has undefined or potentially broken behavior depending on the Ollama version. This mirrors the `trainingToken` guard in `RestClientOcrClient.java:107`.
### 7. OLLAMA_API_KEY behavior in Ollama 0.6.5
### 7. OLLAMA_API_KEY behavior in Ollama 0.6.5 and 0.30.6
**Empirically verified (2026-06-06) on both `0.6.5` and `0.30.6`:** `OLLAMA_API_KEY` does **not** enforce request authentication in either version.
@@ -185,7 +185,7 @@ The model is pre-pulled by `ollama-model-init` before the main service starts (v
**Primary control:** `archiv-net` network isolation. Ollama has no externally exposed port (`expose:` only, not `ports:`). The Caddyfile must not route any path to the Ollama service.
**Defense-in-depth:** `OLLAMA_API_KEY` guards against lateral movement from a compromised backend container.
**Note on `OLLAMA_API_KEY`:** Per §7, `OLLAMA_API_KEY` is not enforced in Ollama 0.6.5 or 0.30.6 and provides no authentication barrier against a compromised backend container. `archiv-net` network isolation is the sole effective security control. The env var is retained for forward compatibility only — do not rely on it for access control.
Both `ollama` and `ollama-model-init` receive the ADR-019 hardening baseline:
@@ -235,5 +235,5 @@ The init container re-pulls the model on next startup.
- **Memory pressure:** OCR + Ollama together consume ~14 GB on a 16 GB host. Running the observability stack simultaneously risks OOM kills. Monitor with `docker stats`.
- **CPU inference latency:** `qwen2.5:7b-instruct-q4_K_M` is chosen for CPU viability, but inference on 8 vCPUs will be noticeably slower than GPU-accelerated alternatives. This is acceptable for the family archive use case (low concurrency, not real-time).
- **Three TBD items** (OLLAMA_API_KEY empty-string behavior, `read_only` feasibility, init container peak RSS) must be resolved before the PR is merged. See Decisions §7, §8, §9.
- All three empirical TBD items from the original issue spec were resolved — see §7 (OLLAMA_API_KEY not enforced), §8 (`read_only: true` works), §9 (peak RSS ~108 MiB).
- Model upgrades require a `docker volume rm` to free old weights before pulling the replacement. Document this in runbook/DEPLOYMENT.md.