Coming soon
The observability stack
you never have to build
Uncloud Hub connects to your Uncloud clusters and gives you a dashboard, metrics, logs, and alerts out of the box. See what's happening in production to catch issues early and fix faster.
Built on Uncloud, the open-source multi-node Docker orchestrator that runs apps across your servers.
Deployed is not done
Deploying is the easy part. Production only starts when the deploy finishes.
- Machines provisioned
-
App built and deployed with
uc deploy - Custom domain and HTTPS
- Services healthy right now?
- Logs you can search when something breaks?
- Alerts before users notice?
- Which deploy caused that error spike?
- Memory or CPU creeping up overnight?
that's what Hub is for
Your cluster, at a glance
Hub provides you with a comprehensive view of your cluster's health and activity.
Services
See what's running, where it's deployed, and whether it's healthy. Across every machine.
Machines
See all nodes, their status, and what's running on each one. No terminal required.
Clusters
Manage multiple clusters from one place. Switch between your dev and production clusters in one click.
Observability stack for builders who'd rather be shipping
We host a Grafana and VictoriaMetrics stack for you so you don't need to build and monitor the monitoring.
CPU, memory, disk, and HTTP request metrics
Monitor resource trends, response times, error rates, and traffic patterns for every machine and service.
Custom app metrics
Expose a Prometheus-style /metrics endpoint
or push OpenTelemetry metrics from your app. Graph signups and failed jobs next to HTTP
requests and CPU.
Slow Postgres queries
Find which queries are dragging down response times. Sorted by impact so you know where to look first, not a wall of query logs.
Searchable service and system logs
Filter by service, level, or time range. Jump from an anomaly in your metrics straight to the logs from that exact minute without switching tools.
Pre-built and custom dashboards
Start with the ready-to-use service and machine dashboards. Tweak them or build your own Grafana dashboards.
Alerts on what affects users
Get notified via Slack, email, or PagerDuty when error rates spike or latency crosses your threshold. Detect, diagnose, and resolve an issue before a user reports it.
API and AI agent access
Metrics and logs are queryable via API and CLI using standard PromQL and LogsQL. Build your own tooling on top, or point an AI agent at an incident and have it investigate the root cause for you.
How it works
Get started in three simple steps.
Connect your cluster
Run uc
hub connect on the
cluster you already have. Its machines and services will appear in the dashboard. Your
deploy flow doesn't change.
Deploy the agent
Deploy the Grafana Alloy agent with
uc
alloy deploy or from
Hub. It configures the required exporters and collectors for you. Metrics and logs will
start flowing.
Monitor everything
Explore your cluster, services, metrics, and logs in one place. See how each deploy behaves in production, and let alerts find you when something needs attention.
Disconnect Hub anytime, no lock-in. Uncloud is Apache 2.0 and fully works without Hub. Your apps run on your servers and deploys keep working from the CLI. Disconnecting only removes Hub's dashboards and stored metrics and logs.
Where Hub goes next
Observability is chapter one. Already on the list:
-
Team access
Share cluster access with your team without handing out SSH keys or root credentials.
-
Push to deploy
Deploy applications from your Git repository automatically.
Your most-needed feature
Something missing from this list? Tell us about your pain and it moves up the queue.
Fair questions
Asked bluntly, answered bluntly.
Do I need Hub to use Uncloud?
No. Uncloud is fully functional on its own, today, from the CLI. Hub adds a web UI and an observability stack on top. If you never connect Hub, you lose nothing that exists now.
Is Hub open source too?
No, and no point pretending otherwise. Uncloud, the orchestrator, is Apache 2.0. Hub is a hosted service built on top.
What data leaves my machines?
Metrics and logs will flow to Hub. Storing, graphing, and searching them is the whole point of a hosted observability stack. Your apps, code, and databases stay on your machines. What exactly gets collected and how you control it is being designed right now, with early access users.
What access does Hub get to my cluster?
The connection model is being designed right now with early access users. The intent is the minimum Hub needs to do its job: telemetry flowing out, and nothing controlling your cluster unless you later opt into features like push to deploy. Uncloud itself never stops working without it.
What happens to my apps if Hub goes down?
Nothing. Your apps run on your machines and Uncloud doesn't depend on Hub to operate. Worst case, you lose the dashboard for a while. Not production.
Why not just run my own Grafana stack?
You can. Grafana, Prometheus, and Loki are excellent, and they'll always work with Uncloud. Hub is for people who want the graphs without adopting a monitoring stack as a second hobby. If you enjoy running one, genuinely, carry on.
When can I use it, and what will it cost?
No public date yet. Hub is in active development and everything on this page is a design preview. Invites go to the early access list as pieces become usable. Pricing isn't decided, and early access users will see it before anyone is charged a cent.
Get in before it's finished
Tell Pasha what you need Hub to do. The roadmap is short enough that one email actually moves it.
No spam, no drip campaign. One person reads these. Or email pasha@uncloud.run directly.