Immich storage tiering is only useful if you can explain what sits where
Immich can make sense for a growing photo and video library, but only if the storage layout is clear. Frequently accessed files can stay on faster storage, while older or less-used media can sit on cheaper storage. That is the basic trade-off: speed for the stuff you use all the time, lower cost for the rest.
For a self-hosted setup, that matters because you keep control of the data rather than handing it to a cloud service and hoping the defaults suit you. The privacy angle is real enough, but it only helps if the rest of the setup is tidy. A messy storage plan just gives you a tidy way to lose track of files.
Backup handling and storage labels need some discipline
Immich can work alongside third-party backup tools such as Restic, which lets you back up the media library without doing everything by hand. That is useful, but it does not remove the need to check what is actually being backed up and where it lands.
The storage template side is also worth using properly. Filename patterns and storage labels make large libraries easier to sort through later. If you have ever tried to find one photo in a pile of badly named exports, you already know why this matters.
S3 lifecycle policies are another part of the picture. They can move files between storage classes based on age and access patterns. In practice, that means older media can be shifted to cheaper storage while still being available when needed.
Metadata stripping is still worth doing
Stripping metadata matters if you care about privacy. GPS coordinates, device details, and other embedded data can ride along in image files unless you take them out. Immich can fit into that workflow, which helps keep control over what gets shared.
Tools such as ImageOptim and online services can remove metadata too. The important bit is not the tool itself, but the habit. If you share images regularly, check what is embedded before you upload them. Some platforms will keep more than you meant to share.
Privacy settings are only useful if you actually use them
Immich gives each user control over profile updates, API key generation, and device management. That is a decent baseline for keeping a library locked down without turning the whole thing into a nuisance.
Machine learning features can be disabled if you do not want them running. Private or protected albums are also available, which means you can share selected albums without opening the whole library. For family use, that is the part that usually matters.
Docker Compose makes deployment manageable
Docker Compose is a sensible way to run Immich because it keeps the deployment in one place. Containers separate the application from the rest of the host, and Compose makes it easier to define storage options and network settings without building everything by hand.
It also makes updates and scaling less painful. If you need to change storage or move to a newer release, you are working from the same configuration instead of trying to remember how you bodged it together six months ago.
Setting it up still needs a plan
Start with the storage you already have and decide which media belongs on faster storage and which can sit elsewhere. Then set up Docker, follow the documentation, and keep the storage and privacy settings matched to what you actually want.
After that, migrate the existing library and test the backup and metadata handling before you trust it. Once the library grows, review the storage policies and privacy settings again. The defaults are not the point; the layout is.
