How to add healthcheck for Bunny without overwhelming logs?

[PoroshkinaVV]

I have an app within k8s. For my probes I want to check if bunny is up and ready to work.

My code for healtcheck:

client = ::Bunny.new(config)
client.start
client.close

Probes starts every 10s and I get really huge amount of info messages, that has no useful information for my app, caused by line client.start.
This leads me to the line:

bunny/lib/bunny/session.rb

Line 1273 in 8e79ee8

@logger.info "Heartbeat interval used (in seconds): #{@heartbeat}"

So I want to ask, is there a possibility to change it's log_level to debug from info?
Or, maybe, there is some best practice, how to add healthcheck for Bunny, without need to start session?

[michaelklishin]

@PoroshkinaVV that message is not particularly important but the only reason why your get "a lot of info messages" is because you have chosen to open new connections with the default log level 6 times a minute per pod.

You can set it to error for those one-off connections without any changes to Bunny or other connection settings.

The Monitoring doc guide for RabbitMQ itself both discusses health checks and recommended monitoring frequency.

[michaelklishin]

@PoroshkinaVV I've reduced the verbosity of that specific message to debug in main. You can override the log level used by these frequent one-off connections in your own code without depending on when the next release drops.

[PoroshkinaVV]

@michaelklishin thank you for your answer!!

I totally understand that.

the only reason why your get "a lot of info messages" is because you have chosen to open new connections with the default log level 6 times a minute per pod.

That's why I asked about alternative ways of healthchecks:

how to add healthcheck for Bunny, without need to start session?

How can I do that?

You can set it to error for those one-off connections without any changes to Bunny or other connection settings.

This guides covers how to monitor rabbitmq itself. But I need to understand, if my application can connect to rabbitmq and don't have access to monitor Rabbitmq itself.

The Monitoring doc guide for RabbitMQ itself both discusses health checks and recommended monitoring frequency.

[michaelklishin]

It makes sense monitor the number of unsuccessful authentication attempts and various operational metrics such as consumer delivery and acknowledgement rates. Bunny does not dictate how you do that.

I don't see much value in connecting every 30-60s, although at that rate, it won't do any harm unless you have tens or hundreds of instance of your app. Connection churn should be avoided.