You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Current ZookeeperLoadBalancerConfiguration implementation caches service instances for 35 seconds by default.
If any instance get down in this period we will load balance on it anyway.
Of course, we can disable service instance cache, but it is not good by performance reasons (and zookeeper possible downtime).
In our project we implemented ServiceInstanceListCacheInvalidator which is subscribed to TreeCache from ZookeeperServiceWatch.
It removes single service instances (one cache entry) from cache when it is notified about that service instances change.
With this feature we can safely increase cache ttl to a relative big value (15 minutes or more).
What do you think about it? Can i make pull request with this feature?
The text was updated successfully, but these errors were encountered:
Current
ZookeeperLoadBalancerConfiguration
implementation caches service instances for 35 seconds by default.If any instance get down in this period we will load balance on it anyway.
Of course, we can disable service instance cache, but it is not good by performance reasons (and zookeeper possible downtime).
In our project we implemented
ServiceInstanceListCacheInvalidator
which is subscribed toTreeCache
fromZookeeperServiceWatch
.It removes single service instances (one cache entry) from cache when it is notified about that service instances change.
With this feature we can safely increase cache ttl to a relative big value (15 minutes or more).
What do you think about it? Can i make pull request with this feature?
The text was updated successfully, but these errors were encountered: