Open Source Dashboard
Dashboard is one of the most overloaded words in software. It can mean a start page that gathers the services you run into one screen of links and live status, an analytics surface that turns a database into shared charts, or a monitoring wall that watches metrics second by second. These are largely unrelated jobs, and the first useful decision is naming which one you are building. Open source helps the same way across all three: your service map lives in a config file you own, your queries and metric definitions stay in readable text instead of a vendor's account, and who may see the numbers never depends on a subscription that can lapse.
Netdata
Sees every metric every second, then uses per-metric ML to flag what looks abnormal
Grafana
Open-source platform for monitoring and observability with dashboards, alerting, and mixed data sources
Metabase
Open source analytics for asking questions, building dashboards, and embedding BI
Glance
Self-hosted dashboard for RSS, social, weather, market, and server feed widgets
Homepage
Highly customizable application dashboard with Docker service discovery and integrations for over 100 services
Redash
Browser-based SQL analytics for querying, visualizing, scheduling, and sharing data
Dashy
Self-hosted homepage for homelabs with status checks, widgets, themes, and a built-in UI editor
OpenObserve
Open source observability platform for logs, metrics, traces, RUM, and pipelines
APITable
API-oriented low-code spreadsheet for collaborative apps, automation, and embedded database views
Our picks
The right pick depends on which kind of dashboard you need, so here is one strong option for each.
Metrics and observability dashboards: Grafana Query, visualize, and alert on metrics and logs wherever they are stored, mixing several data sources in one graph. Dynamic dashboards with template variables and ad-hoc exploration suit time-series and operational data, and alert rules notify Slack, PagerDuty, and similar tools. Best when the data is monitoring signals rather than business tables.
Self-serve BI for non-SQL users: Metabase A visual query builder lets people ask questions and build dashboards without SQL, while analysts drop into the SQL editor for harder work. Dashboards get filters and auto-refresh, teams set alerts and schedule subscriptions to email, Slack, or a webhook, and charts embed into your own product. Best when business users, not just engineers, explore data.
Homelab start page: Homepage A start page for self-hosted services, configured through YAML files or Docker label discovery, with over 100 service integrations and container stats. It is statically generated and proxies backend API requests to keep API keys hidden. It suits one front door to everything you run.
Real-time system monitoring: Netdata Charts system and application metrics at one-second granularity, so short spikes a coarser tool would average away show up immediately, and dashboards populate themselves with little configuration. A model trained per metric flags anomalies, and alerts reach email, Slack, and PagerDuty. Data can stay on your own infrastructure with no central collection.
Deciding which kind of dashboard you actually need
Three unrelated tools hide under the same word here, and telling them apart saves the most time. Homepage, Dashy, Glance, and Homer are start pages that collect the services you self-host into one screen of links, icons, and simple status, usually configured through a YAML file or Docker labels. Metabase, Redash, Grafana, and Lightdash are analytics dashboards that sit on a database or metrics layer and turn queries into shared charts. Netdata and Checkmk are a third thing again, built to watch systems in real time and alert when something breaks. Sort your need into one of these before reading further.
If you came for the analytics kind, the real question is who owns correctness after a chart is published. A code-first tool like Grafana or Lightdash keeps queries and metric definitions in version-controlled text, which makes review and rollback easy but asks more of business users. A point-and-click tool like Metabase lets anyone build a question without SQL, which is faster but lets inconsistent filters and duplicate metrics multiply. Check too that permissions reach the data row and not just the page, and that authentication maps to your identity provider before a dashboard quietly becomes a semi-public app.
For the start-page kind, the trade-offs are smaller but still real. Homer serves plain static files with no backend, which is about as low-maintenance as it gets, while Homarr and Dashy add authentication, drag-and-drop editing, and live widgets at the cost of running a real service. Whichever kind you choose, keep the configuration in files you can back up, because a dashboard that exists only inside an undocumented database is the hardest one to rebuild later.
Related categories
Frequently asked questions
What is the difference between a homelab start page and a BI dashboard?+
They solve unrelated problems. A start page like Homepage, Dashy, or Glance gathers the services you run into one screen of links, icons, and simple status, usually from a YAML file or Docker labels. A BI dashboard like Metabase, Redash, or Grafana connects to a database or metrics layer and turns queries into shared charts and reports. If you want a front door to your self-hosted apps, you want the first kind; to analyze data, the second.
Which dashboards let me build charts without writing SQL?+
Metabase is built around a visual query builder, so anyone can ask questions and build dashboards without SQL, dropping into a SQL editor only when needed. Grist and APITable come at it from a spreadsheet and database angle with charts and views. Grafana and Redash are more SQL-first, aimed at people comfortable writing queries. If nontechnical teammates will author reports, favor the visual-first tools and reserve the code-first ones for engineers.
Can I run one of these on a wall display or TV?+
Yes, though check refresh behavior first. A visual refresh is not the same as fresh data: a panel can redraw every few seconds while still showing a stale cache. Homer and Dashy install as PWAs and suit an always-on start page, and Netdata streams per-second metrics that update live for an operations wall. Test with realistic panel counts and concurrent viewers, because dashboard performance usually fails from many small queries rather than one big one.
How do these dashboards connect to my existing databases?+
Analytics tools connect through database drivers, APIs, or a modeled metrics layer. Redash advertises more than 35 SQL and NoSQL sources including Postgres, BigQuery, and Snowflake, Grafana mixes multiple data sources in one graph, and Lightdash builds its metrics on top of a dbt project. For production, create dedicated read-only service accounts, set query timeouts, and point dashboards at replicas or a warehouse rather than letting them hit transactional systems directly.
Can I embed an open source dashboard in my own app or customer portal?+
Some are built for it. Metabase can embed charts and dashboards in your own product, and Grafana and Redash expose shareable views and APIs. Embedding quality varies, so verify signed URLs, single sign-on, theming, and above all row-level filtering enforced on the server, not through URL parameters. For customer-facing use, confirm tenant boundaries hold server-side and test load time inside your application shell before you rely on it.
Which of these can discover my Docker containers automatically?+
Several start pages read Docker directly instead of making you list every service by hand. Homepage discovers services from Docker labels and can show container stats, Flame picks up apps from container labels, and Glance includes a Docker containers widget. Homarr and Dashy add live status and widgets for self-hosted services too. This is the feature that separates a dashboard that maintains itself from one you edit by hand every time you add a service.
Should I self-host the dashboard or use the vendor's cloud?+
Self-hosting matters most when the dashboard sits close to internal databases, needs private network access, or must keep data in a specific place. It gives you control over connectivity, secrets, and upgrades, and makes you responsible for uptime, backups, and patching. Netdata, Grafana, Metabase, and OpenObserve all offer an optional hosted service if you would rather not run one. Test the deployment model against your slowest data source and strictest access rule.