Do these 3 things before closing this tab:
1Fix the driver behind crashes, sound loss and screen glitches2Repair Windows errors before they cause bigger problems3Scan for outdated or missing drivers - takes under a minuteFor a new Plausible Community Edition server, choose a Docker-capable VPS with a persistent SSD, a CPU that supports SSE 4.2 or NEON, and 4 GB of RAM when Plausible is the main workload. The current Community Edition (CE) README documents 2 GB as the minimum recommendation for running ClickHouse and Plausible without out-of-memory errors, while an older official getting-started guide recommends 4 GB. Treat 2 GB as a small-installation floor and 4 GB as the safer production starting point.
Need a server to try this on? Antihost offers Linux VPS plans with NVMe storage, full root access and a 14-day free trial.
You also need DNS control for the hostname, a correctly formed BASE_URL, a unique SECRET_KEY_BASE of at least 64 bytes, HTTPS, and enough persistent storage for PostgreSQL, ClickHouse, Docker, logs, backups, and the operating system. Plausible publishes no universal vCPU count, disk quota, visitor ceiling, or event-rate guarantee, so traffic and retention—not a single advertised plan size—must determine the final VPS.
As an Amazon Associate I earn from qualifying purchases.
PC Slower Than It Used to Be?
A free scan shows the junk files, broken settings and background clutter dragging Windows down - then fixes them in one click.Free scan · Windows 10 & 11Crashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minuteWhat Plausible needs at a glance
Plausible Community Edition is a Docker-based application composed primarily of the Plausible web service, PostgreSQL, and ClickHouse. PostgreSQL stores application and user data; ClickHouse stores analytics events. Your host must therefore provide both a container runtime and reliable, persistent storage rather than an ephemeral disk.
| Component | Documented requirement or guidance | Practical interpretation |
|---|---|---|
| Container runtime | Docker and Docker Compose are required by the CE repository. | Use a VPS or physical host where you can install and update Docker and Compose. |
| CPU architecture | x86_64 is supported; ARM64 is supported when NEON is available. ClickHouse requires SSE 4.2 or NEON (or a newer equivalent). | Check the actual instruction set, not just the advertised architecture. |
| Memory | CE README v3.2.1 recommends at least 2 GB of RAM. An older official guide recommends 4 GB. | Use 4 GB for a new production VPS when Plausible is the main service; reserve 2 GB for small, closely monitored installations. |
| CPU cores | No fixed official vCPU count is published. | Start with 2 vCPUs for a small installation and add capacity when ingestion, queries, or other services compete for CPU. |
| Disk | No universal disk size is published. | Use persistent SSD storage and size it for event retention, Docker images, logs, backups, and operating-system headroom. |
| Public URL | You control DNS and set BASE_URL to the complete URL, including scheme and hostname. |
Plan the domain or subdomain before starting the containers. |
| TLS | The current CE setup can bind HTTP/HTTPS ports and request Let’s Encrypt certificates. The older stack expects a reverse proxy for SSL termination. | Follow the instructions for your release, or put a correctly configured reverse proxy in front. |
| Secrets | SECRET_KEY_BASE must be random and at least 64 bytes. |
Generate a new value for each installation and keep it out of public repositories. |
The 2 GB and 4 GB figures are version-specific recommendations, not mutually exclusive hard limits. The CE README is the newer source for the 2 GB statement; the older guide remains useful for deployment and operational context.
#1 Best Overall
CPU and architecture requirements
Instruction-set support matters
The CE README states: “CPU must support SSE 4.2 or NEON instruction set or higher (required by ClickHouse).” On x86 systems, SSE 4.2 is normally available on modern 64-bit virtual machines, but low-cost or unusual virtualized hardware can expose a different feature set. On ARM64, NEON support is required. Ask the VPS provider for the instruction flags exposed inside the guest, or verify them on the running host before moving production data.
How many cores?
Plausible does not publish a required vCPU count. A two-vCPU instance is a reasonable starting point for a small site because ClickHouse ingestion, database work, web requests, and background tasks can overlap. More cores help when several sites send events simultaneously, dashboards run expensive queries, retention jobs overlap with ingestion, or other applications share the machine. Increasing cores cannot compensate for insufficient RAM or an undersized disk.
Quick wins for a faster PC:
Repair Windows errors before they cause bigger problemsFix Now →Scan for outdated or missing drivers - takes under a minuteDriver Scan →RAM: is 2 GB enough?
The current CE README (repository version 3.2.1, 2026) recommends at least 2 GB of RAM “for running ClickHouse and Plausible without fear of OOMs.” That is the lowest documented recommendation, not a promise that every traffic pattern will remain healthy on a 2 GB VPS. The older official getting-started guide recommends 4 GB minimum for Community Edition and notes that requirements depend on traffic.
When 2 GB can work
A 2 GB server can be appropriate for evaluation, a very small site, or a low-volume installation with short retention. Keep the host otherwise quiet, monitor memory and swap, and leave room for the operating system, Docker, PostgreSQL, ClickHouse, and temporary jobs. An out-of-memory kill can interrupt more than the web container: ClickHouse or PostgreSQL may be terminated while writing data.
Why 4 GB is the safer starting point
Choose 4 GB when Plausible is intended for production, when several sites share the instance, or when you want headroom for upgrades, migrations, backups, and dashboard queries. The extra memory also reduces the chance that a routine Docker image update or database operation competes directly with ClickHouse for the last available megabytes.
When to move to 6–8 GB or more
Increase memory when event volume grows, retention is extended, dashboard queries become slow, the host runs additional services, or monitoring shows sustained pressure. A 4 vCPU, 6–8 GB configuration is a sensible next tier for larger event volumes or a shared host, but it remains an operational rule of thumb rather than a Plausible capacity guarantee.
Storage and retention planning
No official universal disk quota or exact bytes-per-event figure is published. Your disk must hold ClickHouse event data, PostgreSQL data, Docker images and layers, container logs, the operating system, temporary files, and at least one usable backup set. ClickHouse growth follows event volume and retention, so a site with a modest number of visitors but long retention can consume more space than a busier site with aggressive data expiration.
A practical sizing method
- Estimate the number of page views and custom events per day across every site that will use the instance.
- Decide how long analytics data must remain queryable.
- Add space for PostgreSQL, Docker, logs, upgrades, and temporary maintenance files.
- Reserve headroom rather than filling the volume to its advertised capacity; database performance and recovery become harder when free space is low.
- Choose a provider that allows disk expansion or attachable volumes so growth does not require a disruptive migration.
Use SSD-backed persistent storage. Avoid a VPS whose primary volume is temporary or erased when the instance is rebuilt. Backups should be stored separately from the live disk; a snapshot on the same failure domain is not a complete backup strategy.
DNS, BASE_URL, ports, and HTTPS
DNS and the public hostname
You need authority over the domain or subdomain that will serve Plausible. Create the DNS record before requesting certificates and allow time for propagation. Set BASE_URL to the full canonical address, including https:// and the hostname. Plausible is not designed for installation under a subfolder such as example.com/analytics; use a dedicated hostname instead.
Port and TLS choices
The current CE setup can bind HTTP and HTTPS ports and request Let’s Encrypt certificates. Follow the port variables and certificate flow for the exact release you deploy. The older stack listens on port 8000 and does not terminate SSL itself, so it requires a reverse proxy such as Caddy, Nginx, Traefik, or an equivalent managed proxy in front. Whichever design you choose, expose only the ports you intend to serve, redirect plain HTTP to HTTPS where appropriate, and renew certificates before expiration.
Recommended Free Tools
Rank #3
- New
- Mint Condition
- Dispatch same day for order received before 12 noon
- Guaranteed packaging
- No quibbles returns
Secrets, databases, and email
Generate a unique application secret
Set SECRET_KEY_BASE to a cryptographically random value of at least 64 bytes. The configuration documentation shows an OpenSSL generation method and warns not to reuse its example. Store the value in protected environment configuration or a secrets manager; changing it casually after deployment can invalidate sessions or other signed data.
Understand the database split
PostgreSQL holds application and user information, while ClickHouse holds analytics events. Back up both according to the procedures for your release. A backup of only the web container or only PostgreSQL is not a complete recovery point for an analytics installation.
SMTP and verification
Configure SMTP if your registration or account flow uses email verification. Without working email delivery, new users may be unable to complete verification. The documentation also describes a manual administrator verification procedure, which can help during initial setup or an SMTP outage but should not replace a reliable mail configuration for a production service.
Recommended VPS profiles
The following profiles translate the published 2 GB and 4 GB guidance into starting points. They are operational recommendations, not an official Plausible sizing chart.
| Profile | Starting specification | Suitable for | Watch-outs |
|---|---|---|---|
| Small test instance | 2 vCPU, 2 GB RAM, persistent SSD | Evaluation or a very small site | Monitor ClickHouse, swap, and out-of-memory events; avoid running unrelated services. |
| General small production | 2–4 vCPU, 4 GB RAM, persistent SSD | One or several low-to-moderate traffic sites | Expand disk and RAM as event volume and retention increase. |
| Higher traffic or shared host | 4 or more vCPU, 6–8 GB or more RAM, faster SSD | Larger event volume, longer retention, or multiple applications on one VPS | Benchmark your workload and monitor continuously; there is no official capacity guarantee. |
Installation sequence
- Validate the host. Confirm Docker, Docker Compose, persistent SSD storage, and SSE 4.2 or NEON support. Confirm that you can open the required HTTP/HTTPS ports and install security updates.
- Create the DNS record. Point the chosen domain or subdomain at the VPS and decide whether a reverse proxy will terminate TLS.
- Obtain the CE files. Clone the official Community Edition repository. Pin a documented release when reproducibility matters instead of tracking an unreviewed moving branch.
- Set the URL and secret. Enter the complete
BASE_URLand generate a uniqueSECRET_KEY_BASEof at least 64 bytes. Check for accidental whitespace or a missing URL scheme. - Configure networking. Set the HTTP/HTTPS ports required by your release, or configure the reverse proxy to forward requests to the application’s internal port. Ensure the proxy passes the original scheme and host.
- Start the stack. Run
docker compose up -dfrom the CE directory. Then inspect container status and logs for Plausible, PostgreSQL, and ClickHouse. - Complete first-run checks. Open the HTTPS URL, create the administrator account, and verify that the dashboard loads. If email verification is enabled, test delivery before inviting other users.
- Establish operations. Configure backups for both databases, disk and memory alerts, certificate renewal monitoring, update windows, and a tested restore procedure before sending production traffic.
Monitoring and capacity signals
Watch memory usage, swap activity, CPU saturation, disk utilization, disk latency, ClickHouse and PostgreSQL logs, container restarts, and certificate expiry. Alert before the disk is full rather than after writes fail. Record event volume and retention changes alongside resource graphs so a later increase can be tied to an actual workload change.
A recurring out-of-memory kill indicates that the host is undersized, another service is consuming the budget, or a workload has changed. Sustained high CPU with normal memory points toward more vCPUs or query and retention review. Rapid disk growth calls for checking retention, event instrumentation, logs, and backup placement before simply adding capacity.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Common deployment failures and fixes
ClickHouse will not start
Check CPU features first. A virtual CPU without SSE 4.2 or an ARM64 host without NEON cannot satisfy the documented ClickHouse requirement. Move to a compatible architecture or instance type rather than trying to tune the container.
The Tool Desk
Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →The site loads over HTTP but not HTTPS
Verify DNS resolution, firewall rules, certificate issuance, and the port mapping for your CE release. If using the older stack, confirm that the reverse proxy—not the Plausible container—is responsible for TLS termination and that it forwards to port 8000.
Users cannot verify their email
Check SMTP credentials, sender policy, DNS records for the mail domain, and outbound network access. Use the documented manual administrator verification only as a recovery or bootstrap measure.
The server repeatedly runs out of memory
Inspect which container is being killed, stop unrelated services, reduce unnecessary retention or query load, and move from 2 GB to 4 GB or more when pressure persists. Do not treat a one-vCPU/1 GB anecdote as a benchmark: a support-thread user reported repeated outages at that size and later moved to 4 vCPU and 6 GB, but that experience is not an official requirement.
Best Value
Security, privacy, and operator responsibility
Plausible Community Edition is the free self-hosted release and is open source under AGPLv3 or later. Plausible’s security documentation states that “The raw IP address and User-Agent are never stored in logs, databases or anywhere on disk.” That describes Plausible’s stated handling; you still control reverse-proxy logs, host-level logging, backups, administrator access, and retention of those operational records. Restrict SSH and database access, apply operating-system and container updates, use strong administrator credentials, and encrypt or tightly control backup copies.
The official guide notes that self-hosted operators apply updates themselves. Releases receive changes after they have been battle-tested in the hosted service, but you remain responsible for scheduling upgrades, running any required migrations, checking release notes, and retaining a rollback or restore path.
How to choose between VPS plans
Compare plans on the factors that affect this workload rather than headline CPU counts alone:
- RAM headroom after the operating system and Docker are accounted for.
- CPU architecture and exposed SSE 4.2 or NEON instructions.
- SSD performance, endurance, and the ability to expand the volume.
- Snapshot and off-host backup options.
- Network reliability and the ability to configure DNS, firewalls, and reverse proxies.
- Server region and any data-residency requirement.
- Support quality and the total recurring cost of storage, backups, and transfer.
There is no published maximum visitor count, events-per-second limit, exact disk-consumption formula, or required vCPU number that applies to every deployment. Start conservatively, measure your own event rate and retention, and scale before resource exhaustion affects data collection.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
