-
Notifications
You must be signed in to change notification settings - Fork 45
ADR-42 support failover on overflow policy as well as priority client policy #357
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I think this explains rationale and use case, but is a bit lackluster on the behaviour details itself.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM!
This adds the ability for overflow clients to take over the consumer when no unlimited clients are active. Without this ability the consumer could stall until the limits are reached and once reached again stall once the thresholds fall below the stated limits resulting in unconsumed messages. Essentially the current design works as long as all regions have actively consuming clients, this adds a feature that allows regions to create a cascading HA setup between them where for example us-east could take over for us-west after 5 seconds and eu-west would only step in after 10 seconds Signed-off-by: R.I.Pienaar <[email protected]>
7e3767b
to
1bdc90a
Compare
Would this resolve nats-io/nats-server#5213? |
Thanks @MauriceVanVeen forgot that discussion. I think the concerns there about manual setup overhead and changing conditions still hold true. I think the new named policy approach for prio groups give us a better chance to iterate so we can start manual and later expand to more magical - also means we meet the customer need asap Gor sure good info in the linked issue |
I think |
This adds the ability for overflow clients to take over the consumer when no unlimited clients are active.
Without this ability the consumer could stall until the limits are reached and once reached again stall once the thresholds fall below the stated limits resulting in unconsumed messages.
Essentially the current design works as long as all regions have actively consuming clients, this adds a feature that allows regions to create a cascading HA setup between them where for example us-east could take over for us-west after 5 seconds and eu-west would only step in after 10 seconds