Flask-Caching serves stale values after data source changes

Question

Stale cache after data updateGoal: Ensure that a view decorated with @cached returns up-to-date data immediately after the underlying data source (e.g., a database row) is modified, without waiting for the timeout to expire.Constraint: The Flask-Caching extension does not automatically invalidate cached entries when the data source changes; developers must rely on timeout expiration or manual cache clearing. The stale‑while‑revalidate feature can serve an expired value while refreshing in the background, but it is not enabled by default and requires explicit configuration.Uncertainty: Whether enabling stale‑while‑revalidate with a short refresh window provides acceptable consistency for typical read‑heavy workloads, and how to configure it safely to avoid thundering‑herd regeneration.What timeout and stale‑while‑revalidate values balance freshness and backend load for a frequently updated table?How can a cache‑key prefix be constructed to include request‑specific data (e.g., query strings) without writing a custom callable for every route?Is there a recommended pattern for triggering automatic cache invalidation from SQLAlchemy model change events?

Tasadduq Burney · Answer

Why Flask‑Caching returns stale data
Flask‑Caching does not monitor the underlying data source; a value stays in the cache until its timeout expires or it is explicitly removed.
Confirmed facts

Cache expiration is governed by the timeout argument.
Setting stale_while_revalidate lets the cache serve an expired value while a background refresh occurs.
The cache key can be built from request data (e.g., query string) using make_cache_key or a key_prefix that incorporates request.args.
SQLAlchemy model events (after_insert, after_update, after_delete) can be used to delete the matching cache key.

Steps to get up‑to‑date data after a DB change

Configure the cache with a reasonable timeout and a short stale‑while‑revalidate window, e.g. cache = Cache(app, config={'CACHE_TYPE': 'SimpleCache', 'CACHE_DEFAULT_TIMEOUT': 60, 'CACHE_STALE_WHILE_REVALIDATE': 10}).
Ensure the view’s cache key includes the query string: @cached(key_prefix=lambda: request.full_path.rstrip('?')) or use make_cache_key with request.args.
After committing a change to the model, delete the specific key: cache.delete(make_cache_key(request.full_path.rstrip('?'))) inside an @event.listens_for(Model, 'after_update') listener (and similarly for insert/delete).
If you prefer prefix‑based invalidation, store a version tag in the key and bump the tag on change, then delete all keys matching the prefix.

Missing diagnostic detail
To confirm that the invalidation works across all workers, please specify which cache backend you are using (e.g., SimpleCache, Redis, Memcached). The recommendation changes if you need a distributed delete versus a local one.

Flask-Caching serves stale values after data source changes

Stale cache after data update

1 answer

Why Flask‑Caching returns stale data

Confirmed facts

Steps to get up‑to‑date data after a DB change

Missing diagnostic detail

Your perspective belongs here.

0 question comments