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
first of all thank you for creating this awesome package. You are doing a really great job!
I recently went to production and started evaluating load-balancing strategies and I settled for cluster because it is the cheapest and most efficient I could find.
When I was reading through the docs I first had some issues understanding how exactly cluster works with regards to "load balancing". I think my confusion came from the fact that the environment variable which allows an instance to handle ddp-traffix for other instances is called CLUSTER_BALANCER_URL. Also the textual descriptions refer to the server becoming a balancer by setting the before mentioned environment variable.
As far as I understand each instance automatically becomes a "balancer" when it uses the cluster package and there are instances which declare that they can take ddp-traffic from other instances.
So setting the CLUSTER_BALANCER_URL does not make a node a "balancer" but makes it a "worker" in the distributed balancing scheme.
So every instance is a "balancer", but not every instance has to be a "worker".
Does that make any sense?
If so, I would suggest renaming CLUSTER_BALANCER_URL and the respective js-api to CLUSTER_DDP_WORKER_URL.
The text was updated successfully, but these errors were encountered:
Hi @arunoda,
first of all thank you for creating this awesome package. You are doing a really great job!
I recently went to production and started evaluating load-balancing strategies and I settled for cluster because it is the cheapest and most efficient I could find.
When I was reading through the docs I first had some issues understanding how exactly cluster works with regards to "load balancing". I think my confusion came from the fact that the environment variable which allows an instance to handle ddp-traffix for other instances is called
CLUSTER_BALANCER_URL
. Also the textual descriptions refer to the server becoming a balancer by setting the before mentioned environment variable.As far as I understand each instance automatically becomes a "balancer" when it uses the cluster package and there are instances which declare that they can take ddp-traffic from other instances.
So setting the
CLUSTER_BALANCER_URL
does not make a node a "balancer" but makes it a "worker" in the distributed balancing scheme.So every instance is a "balancer", but not every instance has to be a "worker".
Does that make any sense?
If so, I would suggest renaming
CLUSTER_BALANCER_URL
and the respective js-api toCLUSTER_DDP_WORKER_URL
.The text was updated successfully, but these errors were encountered: