Latest Blog
How Immich Schedules Search Indexing During Large Mobile-Library Imports
Large imports create multiple queued jobs. Search freshness depends on queue progress and shared-resource headroom, not only on upload bandwidth.
How Immich Machine-Learning Indexing Changes a Family Photo Backup Workflow
Immich backup readiness has multiple milestones: copied originals, usable previews, completed indexes, searchable results, and recoverable application state.
Why Does Immich Deployment Architecture Change as Home Servers Add More Services?
More home-server services can turn one Immich host into several operational domains; split roles only when contention or maintenance justifies it.
How to Measure Immich Performance Without Mistaking Cache for Capacity
A capacity test needs representative data, repeatable load, percentiles, and a saturation ruleโnot one fast cached timeline run.
How Much Thumbnail and ML Storage Does Immich Need for Family Photo Backup?
A family Immich library needs measured derivative space plus ML model cache; use per-asset sampling and growth projections before sizing storage.
Why Does Immich Feel Less Responsive Across Different Clients?
Compare Immich clients by separating server latency from client rendering, cached assets, image decoding, and connection-path overhead.
Can Immich Keep Photo Search Usable During Large Mobile Imports?
Usable search needs interactive headroom as well as advancing indexing. Test old queries and newly imported photos as separate household promises.
How Storage Latency Changes Immich Family Photo Backup
High transfer speed can hide slow storage requests. Match Immich user actions to I/O waiting before blaming a disk or changing the storage layout.
