Full-keyspace SCAN in group-flush blocks Redis

• Plugin: Redis Object Cache (free version)
• Redis: 6.2.22, maxmemory 4GB, maxmemory-policy allkeys-lru
• WordPress + WooCommerce (HPOS enabled), PHP 8.4
• Keyspace size at time of testing: ~800,000-900,000 keys total
• Client: PhpRedis 6.3.0

Our store with a large Redis keyspace (800k+ keys), group-based cache invalidation (via the Lua EVAL script that performs a cursor-based SCAN + DEL loop) blocks Redis for 1-1.5 seconds per call. This makes the entire site (and every other visitor) unresponsive for the duration of each EVAL. Operations that trigger many group invalidations in quick succession — most notably WooCommerce’s “Generate variations” bulk action — cause a sustained outage of 60-100+ seconds, during which the site is effectively down for all concurrent users.

Disabled the object-cache.php drop-in (falling back to WP’s non-persistent cache) and repeated the exact same “Generate variations” action on the same product: zero latency issues, instant response, no site slowdown for other visitors. Re-enabled the drop-in and here again the issue.

My question: is there a configuration option to use a more targeted invalidation strategy (e.g. maintaining a per-group key index) instead of a full SCAN on every group flush, in the free version?

Does the Pro version solve this issue?

Thank you in advance