Immich Docker sizing
Immich Docker Hardware Requirements: CPU, RAM and Storage
Docker Compose is Immich’s recommended production installation method, but containerization does not remove the need to size real hardware. CPU, memory, database storage, generated media, video transcoding and machine-learning jobs all run on the host. The useful question is therefore not the bare minimum needed to start the containers, but the hardware tier that remains responsive during imports and background processing.
Quick answer
What this Immich workload needs
Immich currently lists 2 CPU cores minimum, 4 cores recommended, 6GB RAM minimum and 8GB recommended. Treat those as a baseline, not a universal buying recommendation. A new family server usually benefits from more headroom when it will handle video, Smart Search, facial recognition or other Docker services.
Live Amazon hardware
Current products that fit this decision
These listings come from this sprint's dedicated Amazon catalogue. Barebones systems, laptops, enclosures, adapters, unrelated accessories, ambiguous product classes and obvious multipacks are excluded by the normalizer.
Buying decision
Match the hardware to the bottleneck
Size Docker by peak concurrent work. The planner combines users, library size, video intensity and machine-learning features to recommend a practical host tier rather than repeating the minimum requirement.
Interactive decision tool
Immich Docker Host Sizer
Use the workload inputs to identify a practical hardware tier or likely bottleneck before comparing live products. Results are planning guidance; verify current Immich support and the exact hardware configuration before deployment.
Compatibility checklist
Four checks before you purchase
Use a supported deployment
Immich currently favors Linux and Docker Compose for production; full VMs are supported while Docker-in-LXC is not recommended.
Keep PostgreSQL local
Do not treat a faster network as permission to move the standard database data path onto a network share.
Verify hardware access
Quick Sync, NVENC, VAAPI and ML acceleration need the right host, drivers and container device access.
Measure before upgrading
CPU, RAM, SSD, GPU and network bottlenecks create different symptoms. Change the measured bottleneck first.
The official minimum is a starting line
Immich’s requirements currently call for at least 2 CPU cores and 6GB RAM, with 4 cores and 8GB recommended. Those figures prove that an installation can fit within a modest host, but they do not guarantee comfortable performance during a large first import, simultaneous uploads or other containers competing for the same memory.
Docker Compose is the recommended production method
Immich’s installation documentation identifies Docker Compose as the recommended production deployment. That makes complete Linux-compatible mini PCs and desktops natural buying targets. The host still needs adequate storage paths, drivers and permissions; Docker simplifies service packaging, not the underlying hardware limits.
Linux remains the preferred host operating system
Immich recommends Linux or another Unix-like 64-bit OS and warns that non-Linux systems tend to provide a poorer Docker experience. If the server is a dedicated appliance, buying hardware that is well supported by Linux can reduce future troubleshooting compared with selecting a device solely for its Windows desktop features.
Memory pressure shows up during uploads
Immich specifically notes the importance of RAM for smooth uploads. A 6GB minimum leaves little margin for the OS, Docker, filesystem cache and other services. For a fresh purchase, 16GB is a comfortable home-server starting point, while 32GB becomes easier to justify when the machine also hosts databases, home automation, media servers or virtual machines.
CPU work extends beyond transcoding
Even with hardware video acceleration, the CPU handles application logic, metadata extraction, database work and jobs that are not offloaded. A low-power CPU can be enough for a family library but will clear large queues more slowly than a modern Core-class system. Choose based on acceptable processing time rather than demanding workstation performance for every household.
Database storage deserves a local SSD
Immich’s standard Docker Compose configuration says network shares are not supported for the database data location. Put PostgreSQL on reliable local storage and treat the photo library as a separate capacity problem. This separation improves predictability and avoids making UI/database responsiveness dependent on a remote NAS link.
Generated files consume meaningful extra space
Immich notes that thumbnails and transcoded video can increase library size by roughly 10–20% on average. That overhead belongs in storage planning, especially when generated media shares the application SSD. A Docker host with a tiny system drive can run out of space even when the original library lives on large external or network storage.
Hardware acceleration is optional, not magic
Immich can run without GPU acceleration. Hardware transcoding and machine-learning acceleration are performance options that may reduce CPU load or queue time. Add them when the workload benefits; do not turn a supported feature into a mandatory hardware tax for a mostly-photo library with modest usage.
Container count changes the purchase tier
A server running only Immich is easier to size than one also running Jellyfin, Home Assistant, reverse proxies, databases and backup jobs. Count the whole Docker stack. Memory is often the first hidden constraint because every service looks small in isolation until several caches and background jobs become active together.
Plan device access before buying accelerators
Quick Sync, VAAPI, NVENC and machine-learning backends require the relevant devices and drivers to be available inside containers. Before buying a GPU or mini PC, confirm that the intended Linux host exposes the hardware in the way Immich documents. Physical presence of an iGPU is not the same as a working container acceleration path.
A small SSD can become the performance bottleneck
The application drive may hold containers, PostgreSQL, cache, thumbnails, transcodes and logs. If it is nearly full or very slow under sustained writes, adding CPU cores will not fix the bottleneck. Size local SSD capacity with headroom and keep high-value original media protected separately.
Buy for maintainability as well as speed
A good Docker host has accessible firmware updates, replaceable storage, enough RAM, stable Linux networking and cooling that tolerates continuous operation. The fastest mini PC at a given price can be a poor server if it has soldered memory, one tiny SSD slot and no useful recovery path. Server value includes how easily you can maintain it for several years.
Questions people ask
Immich Docker Hardware questions
What are the minimum Immich hardware requirements?
Immich currently lists 2 CPU cores and 6GB RAM minimum, with 4 cores and 8GB RAM recommended.
Is Docker Compose recommended for Immich?
Yes. Immich currently identifies Docker Compose as the recommended production installation method.
Is 8GB RAM enough for Immich Docker?
It can be enough for a dedicated moderate server. More memory is useful for larger imports and multi-service hosts.
Can I run Immich on 4GB RAM?
Immich notes that 4GB systems can run with machine-learning features disabled, but 6GB is the stated minimum for the full standard workload.
Should I run Immich Docker on Windows?
It can run through Windows options, but Immich recommends Linux or another Unix-like 64-bit OS and warns about a poorer Docker experience on non-Linux hosts.
Can the PostgreSQL database live on a network share?
The standard Immich Docker Compose setup says network shares are not supported for the database data location.
How much extra storage do thumbnails need?
Immich says thumbnails and transcoded video can add roughly 10–20% to library size on average. Your mix of media and settings can differ.
Do I need a GPU for Immich Docker?
No. A GPU or iGPU can accelerate supported video or machine-learning jobs, but it is not required for a basic installation.
Is an N100 mini PC enough?
It can fit a light household workload. Choose more CPU and memory for large video libraries, heavy imports or several co-hosted services.
Should I use 16GB or 32GB RAM?
16GB is a strong general home-server starting point. 32GB is easier to justify when Immich shares the machine with many containers or VMs.
Official references and methodology
Verify current Immich support before deployment
Cloudzat distinguishes Immich’s published minimum/recommended requirements from a purchase recommendation. The sizing tool adds workload headroom for uploads, video, machine learning and other services. Live mini PCs are drawn from the sprint’s dedicated Amazon catalogue and must still be checked for Linux support, upgradeability and exact configuration. Cloudzat also separates the host purchase from the bulk-media storage decision so a large photo library does not automatically force an oversized application server.
- Immich - Requirements
- Immich - Hardware Transcoding
- Immich - Hardware-Accelerated Machine Learning
- Immich - Remote Machine Learning
- Immich - Docker Compose Installation
- Immich - System Settings
As an Amazon Associate, Cloudzat may earn from qualifying purchases. Prices, seller terms, exact configurations, firmware and software support can change.