Force a "specific" dns record in order to check every node on a cluster (monitoring DNS RR) #4931
Labels
area:monitor
Everything related to monitors
feature-request
Request for new features to be added
type:enhance-existing
feature wants to enhance existing monitor
π I have found these related issues/pull requests
π·οΈ Feature Request Type
Change to existing monitor
π Feature description
I use a round robin DNS to balance traffic between multiple haproxy instances.
This mean that gateway.mydomain.tld resolves to multiples IP address (ie 123.123.123.123 and 100.100.100.100, etc.),
As it's handled by haproxy, I cannot simply request "tcp:port" on each haproxy because without SNI I cannot check my specific backend (I need to ask haproxy for gateway.mydomain.tld so the request is send to my gateway backend).
If I set gateway.mydomain.tld on kuma, it will request randomly between each "node/ip", meaning that if node 1 is down, it will be detected randomly (and potentially never if I ask for 1 retry before flipping the state of the monitor)
βοΈ Solution
In https monitor, I would like to be able to specific to which is the dns entry must be resolves :
This way I can check every node specifically.
β Alternatives
None.
π Additional Context
No response
The text was updated successfully, but these errors were encountered: