Pre-launch — Gluecron is in final validation. Public signups and git hosting for non-owner users open after launch review.
CodeIssuesDiscussionsWikiPull RequestsProjectsCommitsActionsReleasesContributorsPulse● GatesSecuritySettingsDeploymentsPipelineInsightsAgents✨ Explain✨ Ask AI✨ Workspace✨ Spec✨ Tests▓ Debt Map✨ NL Search🏛 Archaeology
claude/adoring-hopper-5x74bqclaude/affectionate-feynman-ykrf1hclaude/architecture-audit-design-wxprenclaude/build-status-update-3MXsfclaude/charming-meitner-mllb5rclaude/compare-gate-gluecron-s4mFQclaude/confident-faraday-tikcwbclaude/continue-work-XMTlIclaude/crontech-gluecron-deploy-7MIECclaude/crontech-platform-setup-SeKfwclaude/design-2026claude/ecstatic-ptolemy-jMdigclaude/enhance-github-integration-QNHdGclaude/fix-aa-loop-issue-PonMQclaude/fix-actions-and-processclaude/fix-desktop-errors-XqoW8claude/fix-red-workflowsclaude/fix-website-access-6FKJNclaude/gatetest-integration-hardeningclaude/github-audit-improvements-bDFr9claude/gluecron-launch-status-FoMRlclaude/hopeful-lamport-olfCTclaude/issue-to-pr-and-protectionsclaude/jolly-heisenberg-2sg1Qclaude/launch-preparation-QmTb6claude/new-session-xk1l7claude/plan-platform-architecture-kkN4yclaude/platform-analysis-roadmap-1nUGLclaude/platform-launch-assessment-8dWV8claude/polish-platform-release-AeDrUclaude/resume-previous-work-KzyLwclaude/review-crontech-handoff-qYEVqclaude/review-project-completeness-lHhS2claude/review-readme-docs-ulqPKclaude/serene-edison-rj87weclaude/setup-multi-repo-dev-BCwNQclaude/ship-fixes-and-tests-Jvz1cclaude/site-audit-competitive-pctlwgclaude/site-migration-vercel-XstpKclaude/standalone-product-repos-XHFTDcopilot/feat-smart-empty-states-keyboard-first-enhancementcopilot/feat-smart-morning-digest-review-context-restorecopilot/fix-and-process-workflowscopilot/update-ai-powered-code-reviewfeat/debt-mapfeat/push-policy-codeowners-hardeningfeat/smart-digest-contextfeat/stage-impactfeat/t1-secret-migrationfeat/u-polishfeat/w-self-hostfeat/w2-claude-configfix/agent-journey-orphan-sweepgatetest/auto-fix-1776586424172gatetest/auto-fix-1776586534814gatetest/auto-fix-1776590685143gatetest/auto-fix-1776590808199mainops/redeploy-retriggerstyle/dxt-cta-themeworktree-agent-a3377aad30d55da26worktree-agent-a7ef607b7ee1d6c74
auto-update.sh2.9 KB · 66 lines
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
#!/usr/bin/env bash
#
# Fast auto-deploy for the standalone box. Installed as a ~60s systemd timer by
# scripts/standalone-deploy.sh. Polls the deploy branch; when new commits land,
# it pulls, rebuilds, and runs migrations. Push -> live in ~1-2 min, hands-off.
#
# Exits immediately (cheap) when there is nothing new, so a tight interval is fine.
set -euo pipefail

REPO_DIR="/opt/gluecron"
BRANCH="main"
COMPOSE="docker compose -f docker-compose.standalone.yml"

cd "$REPO_DIR"
git fetch origin "$BRANCH" --quiet

local_sha=$(git rev-parse HEAD)
remote_sha=$(git rev-parse "origin/$BRANCH")
[ "$local_sha" = "$remote_sha" ] && exit 0

echo "$(date -Is) deploying $local_sha -> $remote_sha"
git reset --hard "origin/$BRANCH"     # untracked .env / backups are preserved

# NOTE: `|| true` is deliberate. On this Coolify co-tenant box the compose
# also defines a `caddy` service that tries to bind host :80/:443, which
# Coolify's proxy already owns — so `up` reports a non-zero exit for caddy
# even though the app (gluecron) started fine. Under `set -e` that would abort
# the deploy BEFORE the coolify reattach + health gate below. We tolerate the
# partial failure here and instead gate on the app's OWN health further down.
$COMPOSE up -d --build || echo "$(date -Is) compose up returned non-zero (likely the co-tenant caddy port conflict) — continuing to app health gate"
sleep 5

# Co-tenant ingress reattach (Coolify boxes only).
# `docker compose up --build` recreates the gluecron container, which DROPS
# any network attachment not declared in docker-compose.standalone.yml. On
# this box the public ingress is Coolify's Traefik ("coolify-proxy"), which
# reaches the app over the external "coolify" network via the file route
# /traefik/dynamic/gluecron.yaml in the coolify-proxy container. Without this
# reattach, every deploy 502s the site until someone reconnects by hand.
# Idempotent: a no-op if already attached; skipped entirely on a dedicated VPS
# that has no "coolify" network.
if docker network inspect coolify >/dev/null 2>&1; then
  docker network connect coolify gluecron-gluecron-1 2>/dev/null \
    && echo "$(date -Is) reattached gluecron to coolify network" \
    || echo "$(date -Is) coolify network already attached (ok)"
fi

$COMPOSE exec -T gluecron bun run db:migrate || true
docker image prune -f >/dev/null 2>&1 || true

# App health gate — the real success signal (not caddy). Poll the container's
# own /healthz; if the app itself never comes up, exit non-zero so the systemd
# unit records a failed deploy instead of a silent green.
healthy=0
for _ in $(seq 1 20); do
  if docker exec gluecron-gluecron-1 wget -qO- --timeout=4 http://localhost:3000/healthz >/dev/null 2>&1; then
    healthy=1; break
  fi
  sleep 3
done
if [ "$healthy" != "1" ]; then
  echo "$(date -Is) DEPLOY FAILED: app /healthz never came up for $remote_sha" >&2
  exit 1
fi

echo "$(date -Is) deploy complete: $remote_sha (app healthy)"