Optimizing ConfigSync for Large Configurations (1000+ Files)

Most developers start tracking a handful of files — shell config, git settings, editor preferences. But over months and years, the list grows. Add cloud credentials, database tools, multiple programming language configs, project-specific settings, and team shared configs. A power user can easily track hundreds of files. A team sharing configurations can reach into the thousands.

ConfigSync is built for this. Several optimization features work together to keep operations fast regardless of how many files you track.

The single most important optimization. ConfigSync computes a SHA-256 hash of every tracked file and caches the encrypted output. On subsequent pushes, only files whose content has actually changed get re-encrypted and uploaded:

The fast path uses file modification time as a first check. If a file's mtime has not changed, ConfigSync skips it without reading the file or computing a hash. This makes the common case — nothing changed — nearly instant.

By default, a push sends a complete state snapshot to the server. The --changed flag enables delta mode, sending only files that differ from the previous push:

Not every machine needs every config. Filters let you push and pull subsets of your configuration:

A developer tracking 500 files across 20 modules can pull just the 3 modules needed on a cloud VM, transferring 30 files instead of 500. This matters for speed and for minimizing what lands on shared or temporary machines.

Prevent bloat at the source by excluding unnecessary files when you add config directories:

Fewer tracked files means faster hashing, smaller uploads, and cleaner state. Be deliberate about what you track — configuration files belong, logs and caches do not.

Watch mode can become expensive if it monitors too many files. Focus it on the files you actually edit:

The watch filter tells ConfigSync to monitor only the listed files, not your entire tracked set. Combined with delta pushes via push_args, watch mode stays lightweight even when your total configuration set is large. Edit a file, wait five seconds, and only that file gets pushed.

With content-hash caching, delta uploads, targeted filters, and exclude patterns, ConfigSync scales gracefully. The number of tracked files matters for the first push. After that, performance depends on how many files change between operations — and that number is almost always small. A developer with 1000 tracked files experiences the same sub-second push times as someone with 10, because the caching layer makes unchanged files free.