Self-Host Plausible Analytics on Kubernetes

Plausible measures web traffic without cookies and without storing personal data, which removes the consent banner requirement that comes with Google Analytics. The open source Community Edition is the self-hosted build. It needs two databases rather than one, which is the detail most setup guides bury. This page covers the dependencies, the resource footprint, and what changes versus Plausible Cloud. Edka deploys it on a cluster in your own cloud account.

What Plausible Community Edition Needs to Run

Plausible is one application backed by two separate databases:

• PostgreSQL holds accounts, sites, settings, and API keys. Small and transactional.
• ClickHouse holds the analytics events themselves. This is where pageviews land, and it is why Plausible stays fast on large sites.
• A persistent volume for runtime data, 10Gi by default on hcloud-volumes.

Edka points Plausible at existing PostgreSQL and ClickHouse installations already running on your cluster. Host, scheme, port, database, and user are wired automatically, and credentials resolve from Kubernetes Secrets.

Why Plausible Needs ClickHouse

This trips up most self-hosting attempts. Analytics workloads are append heavy and read by aggregate: millions of rows in, a handful of grouped counts out. Row stores like PostgreSQL scan every column to answer that, so dashboards slow down as event volume grows.

ClickHouse is a column store. A query that counts pageviews by path reads only the columns it needs. That is the difference between a dashboard that loads in a moment and one that times out after a few million events.

You cannot drop ClickHouse and run Plausible on PostgreSQL alone. If you want a single database, Umami is the closer fit.

Resource Requirements

Edka’s defaults for the Plausible pod:

Setting	CPU	Memory
Request	250m	512Mi
Limit	1000m	1Gi

Budget separately for the two databases. ClickHouse is the heavier of the pair and benefits from memory more than CPU. On a cluster that already runs PostgreSQL and ClickHouse, Plausible itself is a light addition. Starting fresh, a CPX31 node (4 vCPU, 8GB) is a sensible floor for the three together. See the Hetzner instance types table for the full range.

Community Edition vs Plausible Cloud

The Community Edition is the same analytics engine with a different operating model:

• You run and back up PostgreSQL and ClickHouse.
• Visitor data stays on your infrastructure, which settles data residency questions before they reach a review.
• No per-pageview pricing, so cost tracks your cluster rather than your traffic.
• You control the upgrade schedule by pinning the chart and image version.

Plausible Cloud is the better choice if you do not want to operate two databases. Funding the hosted product also funds the project.

What Edka Provisions

Edka installs Plausible Community Edition from its own Helm chart:

• Plausible application. Deployed with a Recreate strategy, persistent storage, and Prometheus metrics enabled.
• PostgreSQL connection. An existing PostgreSQL installation, database, and user, with optional SSL/TLS verification.
• ClickHouse connection. An existing ClickHouse installation for event storage.
• Gateway access. Optionally expose Plausible through an Envoy Gateway traffic class at a public hostname, used for BASE_URL and origin checks.

A 64 character SECRET_KEY_BASE is generated for you and stays stable for the lifetime of the instance.

Configuration Options

• Registration and access. Invite only, disabled, or open registration, with optional email verification.
• Authentication. Google OAuth credentials and an optional TOTP vault key to encrypt two factor secrets at rest.
• Email delivery. SMTP, Postmark, Mailgun, Mandrill, or SendGrid adapters for transactional email.
• Storage and resources. Volume size, storage class, and CPU and memory requests and limits.
• Placement. Pin pods to a node pool, with automatic tolerations for node pool taints.
• Auto-updates. Automatic chart and image updates with a patch, minor, major, or custom version policy on your chosen check interval.

Deploy Plausible

1. Click “Deploy Plausible Analytics” below.
2. Select an existing PostgreSQL installation, database, and user.
3. Select an existing ClickHouse installation for analytics events.
4. Choose registration mode, storage size, and a public hostname.
5. Deploy and add your first site to get the tracking snippet.

Deploy Plausible Analytics