The best self-hosted analytics tools in 2026 trade Google Analytics’s “free” price tag for full control over your data, at the cost of running your own server. Google Analytics still costs nothing on paper, and that’s exactly the problem. The real price shows up later, in a cookie consent banner nobody clicks through honestly, a data-processing agreement your legal team has to chase down, and a dashboard that increasingly feels built for someone else’s priorities rather than yours.
That’s why self-hosted analytics has gone from a niche privacy stance to a default decision for a lot of developers and product teams in 2026. The catch nobody mentions upfront: “self-hosted” covers everything from a single Go binary that runs in 512 MB to a seven-container stack that wants a dedicated 16 GB server. Picking the wrong one means either outgrowing your tool in three months or running infrastructure three sizes too big for a blog with modest traffic.
This guide ranks nine production-ready self-hosted analytics tools on what they actually track, what self-hosting genuinely requires, and where each one quietly holds features back for its paid cloud tier. If you land on OpenPanel, our complete OpenPanel guide and OpenPanel vs PostHog breakdown go deeper on that specific pick.
Quick Summary: Which Tool Fits Your Situation
| If you want to… | Use this | Best for | Setup time |
|---|---|---|---|
| Track funnels, cohorts, and web analytics in one dashboard | OpenPanel | Product teams replacing Mixpanel | ~10 min (Docker Compose) |
| Get analytics plus flags, replay, and A/B testing in one platform | PostHog | Teams consolidating several product tools | 30–60 min, heavier server |
| Replicate GA4’s full feature depth on your own server | Matomo | Ecommerce, agencies, GA migrations | 30–60 min (LAMP stack) |
| Get the simplest, most polished pageview dashboard | Plausible CE | Marketing sites, blogs, content teams | ~15 min (Docker Compose) |
| Run the lightest possible self-hosted setup | Umami | Solo devs, side projects, small teams | ~10 min (Docker) |
| Track a personal site or blog with almost no server footprint | GoatCounter | Indie hackers, bloggers | ~5 min (single binary) |
| Get funnels and session replay with full self-hosted parity | Rybbit | Startups that want PostHog’s depth without the weight | ~15 min (Docker Compose) |
| Add performance monitoring alongside analytics | Swetrix | Teams that also want Core Web Vitals tracking | ~15 min (Docker Compose) |
| Use a consent-first model where users own their data | Ackee | Public-interest, research, and civic projects | ~15 min (Node.js + MongoDB) |
Best overall for combined web + product analytics: OpenPanel, since self-hosted and cloud ship identical features. Best if you need more than analytics: PostHog, if you’re prepared for the heavier stack it asks for. Best if budget is the only constraint: GoatCounter or Umami, both genuinely free and light to run.
What “Self-Hosted Analytics” Actually Requires
Before the rankings, it’s worth being honest about what self-hosting costs you that a SaaS tool doesn’t.
| Requirement | Why it matters |
|---|---|
| A server with real headroom | Some tools run in 512 MB. Others want 16 GB and a dedicated instance. Undersizing means OOM-killed containers, not just slow dashboards. |
| A domain and HTTPS | Every tool on this list needs its own subdomain and a valid certificate, since the tracking API and the dashboard both need to be reachable over HTTPS. |
| Backups you actually test | Most of these tools split data across two or three databases (commonly PostgreSQL plus ClickHouse). A backup of one without the other leaves you with orphaned data. |
| Someone watching for updates | Self-hosted software doesn’t patch itself. A few tools here (PostHog, Plausible) explicitly warn that security fixes reach Cloud first and self-hosted instances only after you apply the release yourself. |
| Bot filtering, on your own | Cloud versions of several tools here run network-wide bot detection you don’t get on a single self-hosted instance. Expect noisier numbers until you tune filtering yourself. |
How We Ranked These
- Self-hosted feature parity: does the free version match what the paid cloud tier offers, or is it deliberately scoped down?
- What it actually tracks: pageviews only, or funnels, cohorts, and product-level events too?
- Real hardware requirements: sourced from each project’s own documentation, not marketing pages.
- License terms: AGPL, MIT, GPL, and EUPL carry different obligations if you modify and redistribute.
- Deployment complexity: single binary vs. Docker Compose vs. a full LAMP stack.
- Community health: GitHub activity, release cadence, and whether one company funds the whole project.
Pricing and star counts were checked against each project’s official site and GitHub repository as of September 2026; verify current figures before budgeting, since these numbers move.
Master Comparison Table
| Rank | Tool | Category | Self-hosted parity | License | Best for |
|---|---|---|---|---|---|
| 🥇 1 | OpenPanel | Web + product analytics | Full | AGPL-3.0 | Mixpanel replacement, funnels + cohorts |
| 🥈 2 | PostHog | Full product suite | Partial | MIT / ELv2 | Teams wanting flags, replay, and analytics together |
| 🥉 3 | Matomo | Full web analytics suite | Core only (plugins paid) | GPL-3.0 | GA4-depth reporting, ecommerce |
| 4 | Plausible CE | Web analytics | Partial | AGPL-3.0 | Simple, polished pageview dashboards |
| 5 | Umami | Web analytics | Full | MIT | Lightest general-purpose option |
| 6 | Rybbit | Web + product analytics | Full | AGPL-3.0 | PostHog-like depth, lighter stack |
| 7 | GoatCounter | Minimal web analytics | Full | EUPL-1.2 | Personal sites, blogs |
| 8 | Swetrix | Web analytics + performance | Full | AGPL-3.0 | Teams that also want Web Vitals |
| 9 | Ackee | Minimal web analytics | Full (community project) | MIT | Consent-first, transparency-focused setups |
🥇 1. OpenPanel: Best for Combined Web and Product Analytics
OpenPanel is an open-source analytics platform built to combine Mixpanel’s product-analytics depth with the setup simplicity of a tool like Plausible. It tracks pageviews and traffic sources the way a web-analytics tool does, and funnels, cohorts, retention, and user profiles the way a product-analytics tool does, from a single dashboard.
The detail that puts it at the top of this list: self-hosted OpenPanel runs the exact same codebase as its paid cloud tier, with no feature gate separating the two. That’s not universal on this list, as you’ll see below.
Key Features:
- Funnels, cohorts, and retention alongside standard pageview and traffic-source reporting
- Cookieless tracking by default, which typically keeps sites outside cookie-consent requirements
- Session replay with no cap, where tools like Mixpanel limit replay minutes even on paid tiers
- A lean seven-container stack: Next.js dashboard, Fastify API, PostgreSQL, ClickHouse, Redis, and a worker/queue pair
- Real-time dashboard streamed over WebSockets
- SDKs across web, React, Next.js, Vue, Swift, Kotlin, and React Native
How to Deploy:
- Clone the self-hosting branch of the official repository and run the setup script, or
- Deploy it in a few clicks through OpenPanel Hosting on xCloud, which pins the images, provisions private PostgreSQL, ClickHouse, and Redis, and issues HTTPS automatically
Pros and Cons:
| ✅ Pros | ❌ Cons |
|---|---|
| Self-hosted and cloud are feature-identical | Smaller community (6.8k GitHub stars) than Matomo or PostHog |
| Cookieless by default | No published event-volume ceiling for self-hosting the way PostHog publishes one |
| Lighter hardware needs than PostHog or Matomo | Younger project, less third-party documentation |
| Flat, predictable cloud pricing if you ever want that path | Session replay storage still adds up at high traffic |
Best for: Startups and product teams who’ve outgrown a pageview counter and want funnels and cohorts without adopting a much heavier platform.
🥈 2. PostHog: Best for an All-in-One Product Suite
PostHog bundles product analytics with session replay, feature flags, A/B testing, surveys, error tracking, and a data warehouse module, all under one login. It’s the tool to reach for if you’re trying to replace several separate SaaS subscriptions with one self-hosted platform.
The trade-off is real, and PostHog says so in its own documentation: self-hosted deployments are community-supported only, with several cloud-only features (Experimentation, Advanced Paths, Advanced Permissions) held back from the open-source build. PostHog’s own guidance recommends migrating to Cloud once you pass roughly 100,000 events a month.
Key Features:
- Six product modules in one platform: analytics, replay, flags, experiments, surveys, error tracking
- Generous free cloud tier: 1 million events, 5,000 recordings, 1 million flag requests a month, all free
- SQL access for direct queries against your event data
- Active development, with roughly 31.9k GitHub stars and frequent releases
How to Deploy:
- Provision a server meeting PostHog’s stated baseline: 4 vCPU, 16 GB RAM, 30+ GB storage
- Run the official hobby deploy script, which wires up Django, Celery, ClickHouse, Kafka, Zookeeper, PostgreSQL, Redis, MinIO, and Caddy
- Point a domain at it; PostHog issues its own Let’s Encrypt certificate
Pros and Cons:
| ✅ Pros | ❌ Cons |
|---|---|
| Most feature-complete tool on this list | Self-hosted community edition is missing real cloud features |
| Large, active community and fast release cadence | Heaviest hardware requirement here (4 vCPU/16 GB baseline) |
| Generous free cloud tier if you’d rather not self-host | No official CVEs or tagged releases for self-hosted versions |
| Genuinely useful beyond analytics (flags, replay) | Own docs recommend migrating off self-hosting past ~100k events/month |
Best for: Teams that want feature flags, session replay, and analytics in one place and are prepared to run and patch a heavier stack for it. Our full OpenPanel vs PostHog comparison covers this trade-off in more depth.
🥉 3. Matomo: Best for GA4-Level Depth
Matomo (formerly Piwik) has been around since 2007, longer than anything else on this list, and it shows in the feature depth. Ecommerce tracking, a built-in tag manager, visitor-level session logs, and goal funnels are all part of the core, free product, not an add-on.
The catch is that Matomo’s most talked-about features, heatmaps, session recordings, and A/B testing, are sold as paid plugins even on self-hosted instances. The core analytics are genuinely free and complete. The features people associate with a “modern” analytics tool often aren’t.
Key Features:
- Full ecommerce tracking: revenue, product performance, cart abandonment
- Visitor-level logs, closer to old-school Google Analytics than any other tool here
- 54 language translations, the widest localization on this list
- A Google Analytics importer for migrating historical UA (GA3) data
How to Deploy:
- Set up a LAMP stack: PHP 8.1+, MySQL 8+ or MariaDB 10.4+, a web server
- Pull the official Docker image or unpack the tarball onto your server
- Configure the archiving cron job, without which dashboards slow down significantly past roughly 50k pageviews/month
Pros and Cons:
| ✅ Pros | ❌ Cons |
|---|---|
| Most feature-complete free core on this list | Heatmaps, session recordings, and A/B testing cost extra even self-hosted |
| Nearly two decades of development and a large user base | PHP/MySQL stack, more moving parts than a Docker-only tool |
| Strong ecommerce and enterprise track record | Heaviest tracking script here (~23 KB gzipped vs ~1 KB for lighter tools) |
| GA3 data import path for migrations | Needs a beefier server past a few hundred thousand pageviews/month |
Best for: Agencies and ecommerce sites that want GA4-equivalent depth and don’t mind a more involved LAMP-stack deployment.
4. Plausible Community Edition: Best for a Polished, Simple Dashboard
Plausible is deliberately minimal: pageviews, referrers, and goals, presented on a single screen that needs no onboarding. If GA4 feels like flying a plane to check your speedometer, Plausible is closer to the speedometer alone.
Self-hosted Plausible (called Community Edition) runs the same core dashboard as the cloud product, but several features stay cloud-only: marketing funnels, user journeys, ecommerce revenue goals, SSO, the Sites API, and the stronger bot detection the cloud version runs across its whole network. Self-hosted instances rely on basic user-agent filtering alone, which some users report lets noticeably more bot traffic through.
Key Features:
- One-screen dashboard with no configuration required to get useful numbers
- Lightweight script, close to 1 KB
- Three-container stack: the Plausible app, PostgreSQL, and ClickHouse
- WordPress plugin for straightforward integration on WP sites
How to Deploy:
- Provision a server with at least 2 GB RAM and SSE 4.2 CPU support (ClickHouse requires it; virtually all recent x86 VPS instances qualify)
- Clone the community-edition hosting repo and configure
plausible-conf.env - Run
docker compose up
Pros and Cons:
| ✅ Pros | ❌ Cons |
|---|---|
| Cleanest, most approachable dashboard on this list | Funnels, journeys, and revenue tracking are cloud-only |
| Small tracking script, minimal performance impact | Self-hosted bot filtering is noticeably weaker than cloud |
| Straightforward three-container Docker setup | Security fixes reach Cloud first; self-hosted patches lag |
| AGPL-licensed core, genuinely free | Fewer product-analytics features than OpenPanel, PostHog, or Rybbit |
Best for: Marketing sites, blogs, and teams who want traffic numbers without a learning curve and don’t need funnels or user-level tracking.
5. Umami: Best Lightweight General-Purpose Option
Umami sits in the same weight class as Plausible, cookieless, simple, privacy-first, but stays MIT-licensed rather than AGPL and, notably, doesn’t split features between self-hosted and cloud. What you self-host is the whole product.
Key Features:
- MIT license, the most permissive on this list, no copyleft obligations if you fork it
- PostgreSQL or MySQL as the only required database, no ClickHouse dependency
- Lightest resource footprint of any general-purpose tool here, commonly cited around 512 MB
- Custom events and UTM tracking beyond basic pageviews
How to Deploy:
- Provision a small VPS with Docker
- Deploy the official Docker Compose file alongside a PostgreSQL or MySQL instance
- Point your domain at it and add the tracking script
Pros and Cons:
| ✅ Pros | ❌ Cons |
|---|---|
| No self-hosted vs. cloud feature gap | No funnels or product-analytics depth |
| Lightest general-purpose tool on this list | Dashboard is functional, not particularly polished |
| MIT license, no copyleft requirements | Smaller feature set than OpenPanel or Rybbit if you need more than pageviews |
| Fits on the smallest VPS next to your app | Limited team/permission features |
Best for: Developers who want the least amount of server to manage and don’t need product-analytics features.
6. Rybbit: Best PostHog Alternative Without the Weight
Rybbit is the newest tool on this list by a wide margin, and it’s grown fast: roughly 11,000+ GitHub stars since a 2025 launch. It fills a specific gap: product-analytics-style features (funnels, retention, session replay, user journeys) without adopting PostHog’s much heavier infrastructure.
Like OpenPanel, Rybbit ships full feature parity between self-hosted and cloud, which the project calls out directly as a differentiator against Plausible’s cloud-gated features.
Key Features:
- Session replay, funnels, retention, and user journeys, all in the open-source build
- Core Web Vitals monitoring included alongside standard analytics
- 15+ filter dimensions and three-level location tracking (country, region, city)
- ClickHouse-backed, built on Node.js/Koa and React
How to Deploy:
- Provision a Docker-capable server
- Deploy via the official Docker Compose setup, wiring up ClickHouse alongside the app
- Configure your domain and HTTPS
Pros and Cons:
| ✅ Pros | ❌ Cons |
|---|---|
| Full feature parity, self-hosted matches cloud | Youngest project here, less battle-tested at scale |
| Session replay and funnels without PostHog’s stack weight | Smaller community and fewer third-party integrations |
| Active, fast-moving development | Documentation still maturing compared to older tools |
| AGPL-licensed, genuinely free to self-host | No long track record for stability claims yet |
Best for: Startups that want PostHog-style depth (funnels, replay, journeys) without provisioning a 16 GB server for it.
7. GoatCounter: Best for Personal Sites and Blogs
GoatCounter strips analytics down to what a blog or personal project actually needs: pageviews, referrers, and location, delivered by a single Go binary with SQLite as the default database. There’s no real-time dashboard and no funnels, and that’s deliberate.
Key Features:
- ~1 KB tracking script, the lightest on this list
- Single Go binary, no Docker Compose stack required
- SQLite by default, with PostgreSQL available for larger deployments
- EUPL-1.2 license, letting you self-host, modify, and even fork commercially under the same license
How to Deploy:
- Download the compiled binary for your platform, or run the Docker image
- Point it at a domain
- Add the tracking script to your site
Pros and Cons:
| ✅ Pros | ❌ Cons |
|---|---|
| Simplest deployment of any tool on this list | No funnels, no goals dashboard, no real-time view |
| Free self-hosting, no feature gate | Utilitarian interface, hasn’t changed much visually |
| Tiny script, negligible performance impact | Data processes in batches, not live |
| Built by a developer known for accessibility work | Limited team and collaboration features |
Best for: Indie developers and bloggers who want meaningful stats without any operational overhead.
8. Swetrix: Best for Analytics Plus Performance Monitoring
Swetrix covers the same core ground as Plausible or Umami, cookieless pageview and event tracking, and adds something most of this list doesn’t: built-in Core Web Vitals and page-load performance monitoring alongside the analytics dashboard.
Key Features:
- Web performance monitoring (load times, Core Web Vitals) bundled with analytics
- Custom events and user flows, closer to product analytics than Plausible or Umami offer
- Cookieless, GDPR/CCPA-compliant tracking by design
- AGPL-licensed, fully self-hostable
How to Deploy:
- Provision a Docker-capable server
- Deploy the official Docker Compose stack
- Configure your domain and add the tracking snippet
Pros and Cons:
| ✅ Pros | ❌ Cons |
|---|---|
| Performance monitoring most competitors don’t include | Smaller community than Matomo, Plausible, or PostHog |
| Cookieless by design | Hosted plans price on usage, so cost isn’t as predictable at scale |
| Clean dashboard with alerts and exports | Self-hosting still requires standard setup and maintenance work |
| AGPL core, genuinely self-hostable | Less mature ecosystem of guides and integrations |
Best for: Teams that want Core Web Vitals tracking without adding a separate monitoring tool to the stack.
9. Ackee: Best for Consent-First, Transparency-Focused Projects
Ackee stands apart from everything else on this list in one important way: it’s closer to a genuine community project than a single-vendor open-source tool with a paid cloud tier behind it. There’s no company funding it toward a commercial upsell.
It’s minimal by design, tracking basic pageview and event data through a Node.js backend with MongoDB, and its MIT license means no feature gate exists to worry about in the first place.
Key Features:
- Community-maintained, not backed by a single-vendor commercial roadmap
- MIT-licensed, no restrictions on modification or redistribution
- Node.js and MongoDB stack, straightforward for teams already comfortable with that combination
- Minimalist dashboard, focused on essential metrics only
How to Deploy:
- Provision a server with Node.js and MongoDB support
- Clone the official repository and follow the setup guide
- Configure your domain and tracking script
Pros and Cons:
| ✅ Pros | ❌ Cons |
|---|---|
| No commercial vendor behind it, no upsell pressure | Slower development pace than single-vendor tools |
| MIT license, fully permissive | Minimal feature set, no funnels or product analytics |
| Straightforward Node.js/MongoDB stack | Smaller community, less third-party documentation |
| Genuinely transparent, community-governed | Bus-factor risk if core maintainers step back |
Best for: Public-interest, research, and civic projects where a community-governed tool matters more than feature depth.
Cost Breakdown: What You’ll Actually Pay
The software is free across all nine tools when self-hosted. The bill shows up elsewhere.
| Cost category | Typical range | Notes |
|---|---|---|
| Light tools (GoatCounter, Umami) | $5–$10/month VPS | Runs comfortably on the smallest instances most providers offer |
| Mid-weight tools (OpenPanel, Plausible, Rybbit, Swetrix) | $15–$60/month | ClickHouse-backed stacks want more RAM than a single-database tool |
| Heavy tools (PostHog, Matomo at scale) | $35–$150+/month | PostHog’s own baseline is 4 vCPU/16 GB; Matomo climbs similarly past a few hundred thousand pageviews |
| Your time | Hours per month, ongoing | Patching, backups, and bot-filtering tuning are yours unless something manages it for you |
The line item people forget is that last row. A $5/month VPS running a tool that needs weekly attention isn’t actually cheaper than a managed deployment once you count the hours.
Common Mistakes to Avoid
- Undersizing the server for ClickHouse-backed tools. OpenPanel, Plausible, Rybbit, PostHog, and Swetrix all lean on ClickHouse. Check the CPU instruction-set requirement (SSE 4.2 on x86) before you provision, not after a failed deploy.
- Backing up one database and not the others. A ClickHouse-only backup with no matching PostgreSQL backup leaves you with events attached to accounts that no longer exist after a restore.
- Assuming “open source” means “fully featured.” Matomo, Plausible, and PostHog all hold real features back from their self-hosted builds. Check the feature-parity column in the table above before committing.
- Skipping the bot-filtering conversation. Self-hosted instances of several tools here see noisier traffic than their cloud counterparts, since network-wide bot detection doesn’t exist on a single install.
- Picking based on GitHub stars alone. PostHog has more stars than everything else on this list combined, but that doesn’t mean it’s the right fit for a five-person team that just needs funnels.
Which Tool Matches Your Situation
| If you are… | Best choice | Why |
|---|---|---|
| A startup replacing Mixpanel | OpenPanel | Full self-hosted parity, funnels and cohorts included |
| Consolidating flags, replay, and analytics | PostHog | Only tool here covering all three natively |
| An agency running GA4-style reports for clients | Matomo | Deepest reporting, ecommerce, and familiar UI |
| Running a blog or marketing site | Plausible CE or Umami | Simple, lightweight, no product-analytics overhead |
| A solo developer on a tight server budget | GoatCounter | Lightest deployment, genuinely free |
| Wanting PostHog’s depth without PostHog’s weight | Rybbit | Full parity, lighter ClickHouse-based stack |
| Also tracking Core Web Vitals | Swetrix | Performance monitoring bundled in |
| Building for a public-interest or civic project | Ackee | Community-governed, no commercial upsell |
Your 2026 Self-Hosted Analytics Roadmap
Nine tools, and the honest takeaway is that most of them are good at the specific job they were built for. The mistake isn’t picking the “wrong” tool from this list, it’s picking based on feature count instead of matching the tool to what you’re actually going to track and how much server you’re willing to run for it.
What to do this week: write down the three things you actually need to know (funnels? ecommerce revenue? Core Web Vitals? just pageviews?), then match that list against the table above rather than starting from a feature comparison and working backward. Whichever tool you land on, provision the server size its own documentation asks for, not the cheapest tier that technically boots the container.
Whatever you choose, OpenPanel Hosting on xCloud deploys the self-hosted, full-featured OpenPanel build with private PostgreSQL, ClickHouse, and Redis, automatic HTTPS, and coordinated backups already handled. If you’re deploying any of the other tools on this list yourself, our guide to web hosting security best practices covers the basics worth locking down before any of them go live on a public domain, and xCloud’s one-click app catalog covers close to 300 other Docker-hosted apps the same way.
👉 Deploy OpenPanel on xCloud →
If you have found this blog helpful, feel free to subscribe to our blogs for valuable tutorials, guides, knowledge, and tips on web hosting and server management. You can also join our Facebook community to share insights and engage in discussions.
Frequently Asked Questions
What is the best free self-hosted analytics tool?
For general pageview tracking, GoatCounter and Umami are both genuinely free with no feature gate between self-hosted and any paid tier. For product analytics (funnels, cohorts), OpenPanel and Rybbit offer the same full feature parity at a higher feature ceiling.
Which self-hosted analytics tool needs the least server power?
GoatCounter, by a wide margin. It runs as a single Go binary with SQLite, commonly cited as needing well under 512 MB. Umami is close behind, needing only PostgreSQL or MySQL alongside the app itself.
Is self-hosted analytics really GDPR-compliant?
Self-hosting keeps data inside infrastructure you control, which helps with GDPR’s data-transfer rules, but compliance also depends on what you track and how you disclose it. Cookieless tools like OpenPanel, Plausible, Umami, and GoatCounter typically avoid consent-banner requirements; tools using cookies still need proper disclosure regardless of hosting location.
Do I need Docker to self-host analytics?
For most tools on this list (OpenPanel, PostHog, Plausible, Rybbit, Swetrix), yes, Docker Compose is the standard deployment path. Matomo runs on a traditional PHP/MySQL stack without requiring Docker, and GoatCounter ships as a standalone binary that doesn’t need it either.
Which tool has the biggest gap between self-hosted and cloud features?
Matomo holds back the most commonly wanted extras (heatmaps, session recordings, A/B testing) as paid plugins even on self-hosted instances. Plausible CE locks funnels, journeys, revenue tracking, SSO, and its stronger bot detection to the cloud tier. PostHog’s self-hosted edition drops Experimentation, Advanced Paths, and Advanced Permissions.
Can I migrate between these tools later if I pick wrong?
Generally yes, since all nine are open source with exportable data and no proprietary lock-in. The real cost of switching is re-instrumenting your tracking code and rebuilding dashboards and funnels, not a technical barrier to leaving.
Which of these tools is best for a WordPress site?
Matomo has an official WordPress plugin that runs the full application bundled inside WP, which is the most turnkey option if you’re already in the WordPress ecosystem. Plausible also ships a WordPress integration for simpler pageview tracking without Matomo’s added complexity.
How much traffic can a small self-hosted setup actually handle?
It depends heavily on the tool. Matomo’s own guidance suggests a 2 vCPU/4 GB server handles up to roughly 100,000 pageviews a month comfortably. Lighter tools like GoatCounter and Umami handle considerably more traffic on similar or smaller hardware, since they don’t carry ClickHouse’s overhead.