Details: the event_cache_size is used here to create an AsyncLRUCache

synapse/synapse/storage/databases/main/events_worker.py

Lines 242 to 247 in 92c5817

    
           self._get_event_cache: AsyncLruCache[ 
        
               Tuple[str], EventCacheEntry 
        
           ] = AsyncLruCache( 
        
               cache_name="*getEvent*", 
        
               max_size=hs.config.caches.event_cache_size, 
        
           )

Which passes through to LRUCache:

synapse/synapse/util/caches/lrucache.py

Lines 821 to 830 in 41320a0

    
           class AsyncLruCache(Generic[KT, VT]): 
        
               """ 
        
               An asynchronous wrapper around a subset of the LruCache API. 
        
               On its own this doesn't change the behaviour but allows subclasses that 
        
               utilize external cache systems that require await behaviour to be created. 
        
               """ 
        
               def __init__(self, *args, **kwargs):  # type: ignore 
        
                   self._lru_cache: LruCache[KT, VT] = LruCache(*args, **kwargs)

which will scale the max_size by the global factor:

synapse/synapse/util/caches/lrucache.py

Lines 431 to 434 in 41320a0

    
           if apply_cache_factor_from_config: 
        
               self.max_size = int(max_size * cache_config.properties.default_factor_size) 
        
           else: 
        
               self.max_size = int(max_size)

	self._get_event_cache: AsyncLruCache[
	Tuple[str], EventCacheEntry
	] = AsyncLruCache(
	cache_name="getEvent",
	max_size=hs.config.caches.event_cache_size,
	)

	class AsyncLruCache(Generic[KT, VT]):
	"""
	An asynchronous wrapper around a subset of the LruCache API.

	On its own this doesn't change the behaviour but allows subclasses that
	utilize external cache systems that require await behaviour to be created.
	"""

	def __init__(self, args, *kwargs): # type: ignore
	self._lru_cache: LruCache[KT, VT] = LruCache(args, *kwargs)

	if apply_cache_factor_from_config:
	self.max_size = int(max_size * cache_config.properties.default_factor_size)
	else:
	self.max_size = int(max_size)

Uh oh!

documentation incorrectly states that event_cache_size is unaffected by caches.global_factor #13682

Description

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions

documentation incorrectly states that `event_cache_size` is unaffected by `caches.global_factor` #13682