release 1.7.1: host-agent fails when public pages down

2026-07-19 00:52:56 +02:00 · 2026-07-19 00:52:56 +02:00 · 2e3dd6c686
commit 2e3dd6c686
parent 4a4a9cd0ad
5 changed files with 69 additions and 31 deletions
--- a/2
+++ b/2
@ -1 +1 @@
-1.7.0
+1.7.1
--- a/VERSIONS.md
+++ b/VERSIONS.md
@ -5,7 +5,7 @@ Release history for the standalone **taler-monitoring** suite

 ## Version scheme

-Tags and `VERSION` use **MAJOR.FEATURE.FIX** (three components):
+Tags and `VERSION` use **MAJOR.FEATURE.FIX** (three components), introduced in **v1.4.0**:

 | Part | Meaning |
 |------|---------|
@ -13,16 +13,18 @@ Tags and `VERSION` use **MAJOR.FEATURE.FIX** (three components):
 | **FEATURE** | New capabilities (phases, HTML features, i18n packs, …) |
 | **FIX** | Bugfixes and docs that do not add a feature |

-Git tags: `vMAJOR.FEATURE.FIX` (e.g. `v1.7.0`). File `VERSION` omits the `v` prefix.
+Git tags: `vMAJOR.FEATURE.FIX` (e.g. `v1.7.1`). File `VERSION` omits the `v` prefix.

 | Tag | Date (UTC) | Notes |
 |-----|------------|--------|
-| **v1.8.0** | *planned* | **L10n** packs for user-facing strings that need localization: **fr-CH** (FrancPaysan) and **de-CH** (hacktivism / Taler CH German). |
-| **v1.7.0** | 2026-07-18 | **CLI version:** `./taler-monitoring.sh --ver` / `--version` / `-V` prints suite version, git commit, branch, and repo/tag/commit URLs. |
-| **v1.6.0** | 2026-07-18 | **Mail monitoring:** phase `mail` (`check_mail.sh` + `mail-catalog.conf`) for **firefly** (taler.net, gnunet.org) and **pixel**/TSA (taler-systems.com, mail.anastasis.lu); MX/SMTP/IMAP/SPF/DMARC; host-agent `/taler-monitoring-mail*`. |
-| **v1.5.0** | 2026-07-18 | **Mattermost monitoring:** phase `mattermost` (`check_mattermost.sh`) for `mattermost.taler.net` (SPA + `/api/v4/system/ping` + TLS); host-agent HTML `/taler-monitoring-mattermost*`; hourly timer. |
-| **v1.4.1** | 2026-07-18 | **Docs/versioning:** canonical tree `~/src/taler-monitoring`; docs use generic ops/outside-runner wording (no client hostnames); tags always **MAJOR.FEATURE.FIX**. |
-| **v1.3.4** | 2026-07-18 | Summary i18n, monpages code-21, aptdeploy rsync path. |
+| **v1.8.0** | *planned* | **L10n** packs for user-facing strings: **fr-CH** (FrancPaysan) and **de-CH** (hacktivism / Taler CH German). |
+| **v1.7.1** | 2026-07-18 | Host-agent default `STRICT_EXIT=1`; monpages staging-vs-public diagnosis; re-htmlify after public check fails. |
+| **v1.7.0** | 2026-07-18 | CLI `--ver` / `--version` / `-V`: suite version, git commit, repo/tag/commit URLs. |
+| **v1.6.0** | 2026-07-18 | **Mail monitoring:** phase `mail` (firefly + pixel/TSA); MX/SMTP/IMAP/SPF/DMARC; `/taler-monitoring-mail*`. |
+| **v1.5.0** | 2026-07-18 | **Mattermost monitoring:** phase `mattermost` for `mattermost.taler.net`; `/taler-monitoring-mattermost*`. |
+| **v1.4.1** | 2026-07-18 | Generic docs (no client hostnames); canonical install path `~/src/taler-monitoring`. |
+| **v1.4.0** | 2026-07-18 | **MAJOR.FEATURE.FIX** version scheme formalized (three-component tags + `VERSION` file). |
+| **v1.3.4** | 2026-07-18 | Summary i18n, monpages code-21 detection, aptdeploy rsync path. |
 | **v1.3.3** | 2026-07-18 | GIT notes vs GUI notes; monpages in TESTS. |
 | **v1.3.2** | 2026-07-18 | Suite paths only `~/src/taler-monitoring` (no admin-log suite tree). |
 | **v1.3.1** | 2026-07-18 | monpages FQDN check + deploy no longer silent-skip. |
@ -31,33 +33,23 @@ Git tags: `vMAJOR.FEATURE.FIX` (e.g. `v1.7.0`). File `VERSION` omits the `v` pre
 | **v1.1.0** | 2026-07-18 | i18n chrome: English default; French for FrancPaysan. |
 | **v1.0.0** | 2026-07-18 | Initial standalone release. |

-Older short tags (`v1`, `v1.0`, `v1.1`, `v1.2`, `v1.3`) are obsolete; use the three-component tags above.
-
 ## Sticky bar version (v1.3.0+)

-Generated HTML shows the installed release tag (e.g. `v1.7.0`) next to the host line.  
-Clicking opens the tag page on Forgejo:
+Generated HTML shows the installed release tag (e.g. `v1.7.1`) and links to Forgejo:

 `https://git.hacktivism.ch/hernani/taler-monitoring/src/tag/<tag>`

-Resolved via `git describe --tags` after `update-suite.sh`, with fallback to the `VERSION` file.
+Resolved via `git describe --tags` / `VERSION`.

 ## Public monitoring pages (v1.3.1+)

-Host-agent must publish under `DEPLOY_WWW_ROOT` **and** the reverse proxy must `handle` the paths.
-If not, public FQDNs return merchant **JSON code 21**.
+Publish under `DEPLOY_WWW_ROOT` and Caddy `handle`s. Missing handles → merchant **JSON code 21**.  
+See koopa-admin-log issue **I-09** / `apply-monitoring-live.sh`.

 | Check | How |
 |-------|-----|
-| Phase | `./taler-monitoring.sh monpages` (or `pages`) |
-| Auto | host-agent default phases include `monpages`; post-check after HTML write |
-| URLs | `https://$MON_HOSTS$HTML_URL_OK` |
-
-On the stack host (e.g. koopa), publish + Caddy handles:
-
-```bash
-sudo ~/koopa-caddy/apply-monitoring-live.sh
-```
+| Phase | `./taler-monitoring.sh monpages` |
+| CLI version | `./taler-monitoring.sh --ver` |

 ## Language (v1.2.0+)

@ -67,11 +59,11 @@ sudo ~/koopa-caddy/apply-monitoring-live.sh
 | Env | `TALER_MON_LANG=en` or `fr` |
 | Auto | `*lefrancpaysan*` → fr, else en |

-Full locale packs (fr-CH / de-CH) are scheduled for **v1.8.0**.
+Full locale packs (fr-CH / de-CH): **v1.8.0** (planned).

 ## Pin a release

 ```bash
 git clone https://git.hacktivism.ch/hernani/taler-monitoring.git ~/src/taler-monitoring
-cd ~/src/taler-monitoring && git checkout v1.7.0
+cd ~/src/taler-monitoring && git checkout v1.7.1
 ```
--- a/check_monitoring_pages.sh
+++ b/check_monitoring_pages.sh
@ -11,6 +11,8 @@
 #   MONITORING_PAGE_URLS  optional full URL list (overrides host+path construction)
 #   MONPAGES_EXTRA_URLS   additional absolute URLs (space-separated)
 #   TIMEOUT               curl timeout (from lib.sh)
+#   MONPAGES_STAGING_BASE / HTML_OUT  optional staging root for diagnosis
+#   HTML_OK_DIR           staging subdir (default monitoring)
 #
 set -euo pipefail
 ROOT=$(cd "$(dirname "$0")" && pwd)
@ -171,8 +173,25 @@ for u in "${URLS[@]}"; do
 done

 if [ "$ec" -ne 0 ]; then
-  echo "  hint: HTML may exist only under ~/monitoring-sites-staging; public needs"
-  echo "        writable DEPLOY_WWW_ROOT + Caddy handles (sudo ~/koopa-caddy/apply-monitoring-live.sh)"
+  echo "  hint: public pages down (merchant JSON code 21 = Caddy has no /monitoring* handles)"
+  echo "        fix on stack host: sudo ~/koopa-caddy/apply-monitoring-live.sh"
+  # Staging vs public diagnosis (host-agent sets MONPAGES_STAGING_BASE / HTML_OUT)
+  _stg="${MONPAGES_STAGING_BASE:-${HTML_OUT:-${HTML_BASE:-$HOME/monitoring-sites-staging}}}"
+  _ok_dir="${HTML_OK_DIR:-monitoring}"
+  if [ -n "${MON_HOSTS:-}" ] && [ -n "$_stg" ]; then
+    for _h in $MON_HOSTS; do
+      _f="$_stg/$_h/${_ok_dir}/index.html"
+      if [ -f "$_f" ]; then
+        echo "        staging OK: $_f ($(wc -c <"$_f" | tr -d " ") bytes) — not published/served"
+      else
+        echo "        staging missing: $_f"
+      fi
+    done
+  fi
+  _www="${DEPLOY_WWW_ROOT:-/var/www/monitoring-sites}"
+  if [ -e "$_www" ] && [ ! -w "$_www" ]; then
+    echo "        DEPLOY_WWW_ROOT=$_www not writable by $(id -un 2>/dev/null || echo user)"
+  fi
 fi

 exit "$ec"
--- a/host-agent/README.md
+++ b/host-agent/README.md
@ -185,3 +185,10 @@ zypper in qrencode zbar
 Without them, monitoring still runs; QR group WARNs and skips encode/decode.

 Caddy `/var/www/monitoring-sites` is separate (you configured hernani write access).
+
+
+## Exit codes (v1.7.1+)
+
+Host-agent jobs default to **`STRICT_EXIT=1`**: non-zero exit if suite checks fail, `DEPLOY_WWW_ROOT` is not writable, or public monpages (FQDN) fail.
+
+Set `STRICT_EXIT=0` only if you want timers to stay green while still logging errors (not recommended for production).
--- a/host-agent/run-host-report.sh
+++ b/host-agent/run-host-report.sh
@ -9,6 +9,7 @@
 #   4) always write HTML (first run / empty staging too) + Forgejo commit link
 #   5) rsync to DEPLOY_WWW_ROOT (ERROR if not writable — v1.3.1)
 #   6) monpages post-check: public FQDN must serve monitoring HTML
+#   7) exit code: STRICT_EXIT default 1 (v1.7.1) — unit fails if publish/public pages fail
 #
 # Configure via env file (not in git):
 #   ${XDG_CONFIG_HOME:-$HOME/.config}/taler-monitoring/env
@ -412,19 +413,38 @@ else
 fi

 # Post-publish: verify this job's pages via FQDN (same paths as monpages phase)
+_monpages_failed=0
 if [ "${MONPAGES_POSTCHECK:-1}" = "1" ] && [ -x "$ROOT/check_monitoring_pages.sh" ]; then
  echo "--- monpages post-check (public FQDN) ---"
-  export MON_HOSTS HTML_URL_OK HTML_URL_ERR TALER_DOMAIN
+  export MON_HOSTS HTML_URL_OK HTML_URL_ERR TALER_DOMAIN HTML_OUT HTML_BASE
+  export MONPAGES_STAGING_BASE="${HTML_BASE}"
  if ! bash "$ROOT/check_monitoring_pages.sh"; then
    echo "ERROR: public monitoring page(s) missing or not suite HTML via FQDN" >&2
+    echo "       staging HTML_OUT=${HTML_BASE} · DEPLOY_WWW=${DEPLOY_WWW:-}" >&2
    ec=1
+    _monpages_failed=1
  fi
 fi

+# If public check failed after first HTML write, rewrite pages so sticky bar shows monpages ERRORs
+if [ "$_monpages_failed" = "1" ]; then
+  echo "--- re-htmlify after monpages failure (include public outage in report) ---"
+  # append marker to log so converter sees ERROR lines if monpages was post-only
+  {
+    echo ""
+    echo "ERROR monpages: public FQDN monitoring HTML missing or merchant JSON code 21"
+    echo "ERROR monpages: fix with sudo ~/koopa-caddy/apply-monitoring-live.sh (www + Caddy handles)"
+  } >>"$LOG"
+  for host in $MON_HOSTS; do
+    htmlify_host "$host"
+  done
+fi
+
 ls -1t "$LOG_DIR"/run-*.log 2>/dev/null | tail -n +30 | xargs rm -f 2>/dev/null || true

 echo "======== done ec=$ec · log=$LOG · ${COMMIT_URL:-} ========"
-if [ "${STRICT_EXIT:-0}" = "1" ]; then
+# v1.7.1: default STRICT_EXIT=1 so systemd/timer reflects publish/public failures
+if [ "${STRICT_EXIT:-1}" = "1" ]; then
  exit "$ec"
 fi
 exit 0
 @ -1 +1 @@
 .7.0
 .7.1