Connection Recovery is triggered without the old connection being closed

[JanEggers]

Describe the bug

I recently updated 6.8.1 to 7.0.0 and now im seeing massive reconnects without the old connection beeing closed in the broker.
In normal operations we use about 200 connections that are statically allocated so connection churn should be 0. when I look at the dashboard I see massive number of connections and growing.

I see this behavior after some time of running the app in a testenvironment but I was not yet able to narrow down the cause of the behavior. I hooked into the events on connection and see logs that connections are recovered. but I dont see any indication why reconnect was initiated.

Reproduction steps

x

Expected behavior

connection recovery should not inflate the number of open connections.

Additional context

No response

[JanEggers]

looks like connection recovery was just a sideeffect of the many connections created by AspNetCore.HealthChecks.Rabbitmq will continue to investigate

[JanEggers]

confirmed they dont close the connection sry for bothering you

[lukebakken]

sry for bothering you

No problem, thanks for following up here.

It sounds like the AspNetCore.HealthChecks.Rabbitmq could use some work.

I wonder if you hit this? Xabaril/AspNetCore.Diagnostics.HealthChecks#2343

[JanEggers]

https://github.com/Xabaril/AspNetCore.Diagnostics.HealthChecks/blob/6f0275257fde28a4d0e30808fd21ebf661fe9177/src/HealthChecks.Rabbitmq/RabbitMQHealthCheck.cs#L31-L35

basically the code just works for the singleton case in case the factory is used it leaks connections im in the process of creating a bug there