**Applied:** `scripts/patch-nginx-explorer-config.sh` added locations to HTTP and HTTPS blocks. Config deployed via `pct push`.
**Original cause:** `location = /api/config/token-list` and `location = /api/config/networks` in `fix-nginx-conflicts-vmid5000.sh` are not in the live nginx config. Requests fall through to `location /api/` and get proxied to Blockscout.
**Steps:**
1. SSH to Proxmox host or enter VMID 5000.
2. Run inside VMID 5000:
```bash
# From repo root, copy script into container and run:
## 2. Token-Aggregation Service — ✅ FIXED (2026-02-28)
| Issue | Status | Fix |
|-------|--------|-----|
| `/health` returns `{"status":"unhealthy","error":"database \"token_aggregation\" does not exist"}` | ✅ Fixed | DB created; migrations run; service restarted |
**Applied:** Created `token_aggregation` DB; ran migrations; restarted service. Health now returns `"status":"healthy"`.
**Original cause:** The deployed token-aggregation service (port 3001) uses `DATABASE_URL` pointing to a database named `token_aggregation`, but that database does not exist or migrations were not run.
**Steps:**
1. On VMID 5000 (or wherever PostgreSQL runs):
```bash
# Create database if using separate DB:
createdb -U postgres token_aggregation
# Or ensure DATABASE_URL uses explorer_db (migrations create tables there)
## 3. Nginx Proxy Order — Token-Aggregation vs Blockscout
| Issue | Status | Fix |
|-------|--------|-----|
| `/api/v1/` may proxy to Blockscout instead of token-aggregation | ✅ Addressed | `fix-nginx-conflicts-vmid5000.sh` defines `location /api/v1/`**before**`location /api/` (lines 132–142 before 159). When applying config on VMID 5000, use this script to preserve order. |
**Reference:** `explorer-monorepo/scripts/fix-nginx-conflicts-vmid5000.sh` — correct order is in repo; operator should use this script when (re)applying nginx config.
---
## 4. Explorer Homepage / Wallet Page Tests — Intermittent
| Issue | Status | Fix |
|-------|--------|-----|
| `verify-all-systems.sh` "Explorer homepage" or "Wallet page" may fail | ✅ Improved | Timeout for Explorer homepage and Wallet page increased from 15s to 25s in `scripts/verify-all-systems.sh` to reduce failures on slow networks. |
**Note:** Homepage uses `SolaceScanScout|Blockscout|blockscout|<!DOCTYPE`; wallet uses `Chain 138|ChainID 138|Add Chain`. test_endpoint already captures curl output to a variable before grep.
---
## 5. Wallet Page — Grep Pattern
| Issue | Status | Fix |
|-------|--------|-----|
| Wallet page test looked for "Add Chain 138" | ✅ Fixed | Updated to `Chain 138|ChainID 138|Add Chain` |
---
## 6. Token-Aggregation Health Test — Resilience
| Issue | Status | Fix |
|-------|--------|-----|
| Health test expected `healthy` only; service returns `unhealthy` when DB missing | ✅ Fixed | Test now accepts `healthy`, `"status"`, or `unhealthy` (service reachable) |
---
## 7. Token List Validation — CI
| Issue | Status | Fix |
|-------|--------|-----|
| Token lists not validated in CI | ✅ Fixed | `validate-config-files.sh` now validates dbis-138, cronos, all-mainnet, DUAL_CHAIN |
| Workflow not triggered on token list changes | ✅ Fixed | `validate-config.yml` paths include `token-lists/**` and `explorer-monorepo/backend/api/rest/config/metamask/**` |
---
## 8. DUAL_CHAIN Config Sync
| Issue | Status | Fix |
|-------|--------|-----|
| DUAL_CHAIN files in 3 locations could diverge | ✅ Fixed | `scripts/sync-dual-chain-configs.sh` syncs from `explorer-monorepo/backend/api/rest/config/metamask/` to docs and metamask-integration |
**Usage:** After editing DUAL_CHAIN files, run `./scripts/sync-dual-chain-configs.sh` from repo root.
---
## 9. Optional / Future Improvements
| Item | Priority | Notes |
|------|----------|-------|
| Shellcheck in CI | Low | `run-shellcheck.sh --optional` — install shellcheck if desired |
| Token-aggregation DB naming | Medium | Align DATABASE_URL: use `explorer_db` or create `token_aggregation` and document |
| Explorer homepage test timeout | Done (2026-03-02) | Increased to 25s for Explorer homepage and Wallet page in `verify-all-systems.sh` |
When running `verify-backend-vms.sh`, `verify-all-systems.sh`, or NPMplus checks, the following responses are **often expected** and do not necessarily indicate a failure:
| Code | Meaning | Typical cause |
|------|--------|----------------|
| **301** | Redirect | HTTPS redirect (e.g. nginx on :80 redirecting to HTTPS). Service is up. |
| **404** | Not found | Wrong port or path used in the check; or NPMplus/proxy returns 404 for a bare path. Service may still be healthy. |
| **000** | No response | Connection failed from the host running the script: wrong host (e.g. checking NPMplus admin from off-LAN), firewall, or service bound to localhost only (e.g. NPMplus admin on :81 inside CT). |
**Summary:** 301 = HTTPS redirect (normal). 404 = incorrect port/path or NPMplus behaviour. 000 = connectivity/context (host, TLS, or port). Treat as failures only when the intended endpoint and client context match.
1.**Apply nginx fix and deploy config on VMID 5000:** Run `./scripts/apply-remaining-operator-fixes.sh` from repo root (LAN/operator). **Applied 2026-03-02:** nginx fix and explorer config deploy completed successfully.
2.**Token-aggregation DB:** Run `./scripts/apply-token-aggregation-fix.sh` to create DB and run migrations inside VMID 5000. If the container has no `postgres` user, run `createdb`/migrations on the host where PostgreSQL runs, or point token-aggregation `DATABASE_URL` to `explorer_db` and run migrations there (see §2).
3. Restart token-aggregation after DB fix (script does this when postgres user exists).
