Document final local launch verification
This commit is contained in:
parent
3238a9feb3
commit
14d14bf4a0
|
|
@ -3524,3 +3524,40 @@ promoted.
|
||||||
`https://whoneedhelp.imalto.site/healthz/ready` returned the exact ready
|
`https://whoneedhelp.imalto.site/healthz/ready` returned the exact ready
|
||||||
payload after verification. Production, the frozen hackathon test, shared
|
payload after verification. Production, the frozen hackathon test, shared
|
||||||
Caddy, and the public Git remote were not changed or pushed.
|
Caddy, and the public Git remote were not changed or pushed.
|
||||||
|
|
||||||
|
# 2026-08-25 staff-directory and final local quality recheck
|
||||||
|
|
||||||
|
- Read-only inspection of the development database found the two deployed GIN
|
||||||
|
trigram indexes used by the staff user directory
|
||||||
|
(`users_email_trigram_index` and `users_display_name_trigram_index`) plus the
|
||||||
|
cursor-order index (`users_moderation_cursor_index`). With sequential scans
|
||||||
|
disabled only for an `EXPLAIN` proof, PostgreSQL selected both trigram indexes
|
||||||
|
through a `BitmapOr`. On the current 1,016-user development data set the
|
||||||
|
normal planner selected a sequential scan because its estimated cost was
|
||||||
|
lower; this observation is not a million-user performance claim.
|
||||||
|
- The authenticated staff workspace was exercised read-only at desktop
|
||||||
|
1280x800 and mobile 390x844 viewports. Overview, cursor-paginated user search
|
||||||
|
and detail, support/legal, moderation, analytics, and audit-log screens
|
||||||
|
rendered without current browser-console errors. Searching for the existing
|
||||||
|
`SimpleTest` account returned exactly that account. No user status, role,
|
||||||
|
support case, moderation record, or audit record was changed.
|
||||||
|
- The first isolated quality attempt correctly failed ShellCheck `SC2016` on an
|
||||||
|
intentional container-side expansion in the rolling verifier. The command is
|
||||||
|
now locally documented with a scoped ShellCheck suppression matching the
|
||||||
|
existing container-side expansion pattern; no runtime behaviour changed.
|
||||||
|
- The complete replacement unit
|
||||||
|
`codex-heavy-wnh-quality-20260825-r2-20260825-205940-1130636.service`
|
||||||
|
completed successfully in 2 minutes 46.905 seconds with a 281 MiB memory
|
||||||
|
peak and zero swap use. ExUnit reported 487 passing tests. All configured
|
||||||
|
compiler, formatting, xref, Credo, Sobelow, Dialyzer, dependency, Compose,
|
||||||
|
Helm, migration, rollback, observability, infrastructure-image, and final
|
||||||
|
release-image gates passed. The final Debian 13.6 release image reported zero
|
||||||
|
detected vulnerabilities at the configured scan severity.
|
||||||
|
- Exact post-run inspection found no container, network, volume, or temporary
|
||||||
|
audit/security image with the run identifiers
|
||||||
|
`wnh_quality_202608251759401130640` or
|
||||||
|
`20260825175940-1130640`. Both development web replicas had no matched error,
|
||||||
|
exception, crash, failure, or timeout log entry during the inspected preceding
|
||||||
|
hour. This was local verification only: production, the frozen hackathon
|
||||||
|
deployment, shared Caddy, and the public Git remote were not changed or
|
||||||
|
pushed.
|
||||||
|
|
|
||||||
|
|
@ -161,6 +161,8 @@ web_readiness_snapshot() {
|
||||||
: >"$jsonl_file"
|
: >"$jsonl_file"
|
||||||
|
|
||||||
for pod in "${web_pods[@]}"; do
|
for pod in "${web_pods[@]}"; do
|
||||||
|
# PORT is intentionally expanded inside each web container.
|
||||||
|
# shellcheck disable=SC2016
|
||||||
if ! body=$(
|
if ! body=$(
|
||||||
"${kube[@]}" exec -c web "$pod" -- sh -c \
|
"${kube[@]}" exec -c web "$pod" -- sh -c \
|
||||||
'curl --silent --show-error --fail --max-time 2 \
|
'curl --silent --show-error --fail --max-time 2 \
|
||||||
|
|
|
||||||
Loading…
Reference in New Issue
Block a user