feat: add OneUptime template

[Siumauricio]

What

Adds OneUptime (https://github.com/OneUptime/oneuptime) as a template — an open-source observability platform: uptime monitoring, status pages, incident management, on-call rotations, and logs/traces/metrics.

Closes #514

Notes on the architecture

OneUptime consolidated its former ~15 microservices into a single app image, so the stack is now 6 services, faithful to upstream's current docker-compose.yml (pinned to 11.5.6):

Service	Image	Purpose
oneuptime	oneuptime/app	API, dashboard, accounts, status pages, workers, telemetry ingest
ingress	oneuptime/nginx	Entry point (domain attaches here, port 7849). Does essential path rewrites (/identity → /api/identity, etc.) and custom-domain cert plumbing
probe	oneuptime/probe	Executes monitors; self-registers via REGISTER_PROBE_KEY through http://ingress:7849 (host ingress is part of upstream nginx's server_name)
postgres (15), redis (7.0.12), clickhouse (25.7)		Datastores

Details worth calling out:

• ClickHouse needs upstream's two config drop-ins (single-node cluster + embedded Keeper, and Distributed-insert batching) because OneUptime's analytics schema is always ReplicatedMergeTree + Distributed. They're provided via [[config.mounts]].
• SERVER_NAMES_HASH_BUCKET_SIZE=128 on the ingress: generated Dokploy domains (e.g. long *.sslip.io hosts) overflow nginx's default bucket size of 64 and put the ingress in a restart loop.
• The marketing home service is not deployed — upstream's nginx only routes to it when BILLING_ENABLED=true.
• HTTP_PROTOCOL defaults to http; users should flip it to https once TLS is enabled on the domain.
• Resource note: upstream docs list 8 GB RAM / 4 cores as the minimal ("homelab") sizing; observed idle footprint of this stack is ~4.5 GB.

How it was tested

Local (Docker, clean volumes) and deployed on a Dokploy instance via template import:

• All 6 containers healthy/stable (no restarts after 10+ min).
• /, /accounts, /dashboard, /status → 200 through the generated domain.
• Full signup + login E2E against /identity/* (first user becomes master admin).
• ClickHouse analytics schema created (27 tables) on the embedded-Keeper single-node cluster.
• Probe registered and actively polling /probe-ingest/monitor/list (200s in ingress access log).

🤖 Generated with Claude Code

[github-actions]

built with Refined Cloudflare Pages Action

⚡ Cloudflare Pages Deployment

Name	Status	Preview	Last Commit
templates	✅ Ready (View Log)	Visit Preview	6683685