-
Notifications
You must be signed in to change notification settings - Fork 1.7k
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
Vector exits after start when AMQP sink is unhealthy despite setting --require-healthy flag to false #19466
Comments
Hi. I get the same behavior. |
That's hilarious. We are experiencing similar issues with blocking behavior in
This log with I suggest this issue should be replicated in (integration) tests. For now I was unable to find out if a root cause is in lapin. |
Hey @fedorkanin—maybe I can help you with the Go thing, I also run prod stuff w/ Go using Rabbit, and I wrap my connection with this: github.com/assembla/cony. It handles the reconnection problem amqpCli := cony.NewClient(
cony.URL(os.Getenv("RABBITMQ_URL")),
cony.Backoff(cony.DefaultBackoff),
)
queue := &cony.Queue{Name: queueName}
amqpCli.Declare([]cony.Declaration{
cony.DeclareBinding(cony.Binding{
Key: "",
Queue: queue,
Exchange: cony.Exchange{
Name: "amq.direct",
Kind: "direct",
},
}),
})
cns := cony.NewConsumer(queue, cony.Qos(10000))
amqpCli.Consume(cns)
for amqpCli.Loop() {
select {
case message := <-cns.Deliveries():
{
ch <- message
}
// ... |
@andreyvital thank you, but unfortunately cony seems to perform the same blocking call causing the issue – you have just not encountered it yet :) Nevertheless, if I am wrong, please mention the solution in the relevant amqp019-go issues, this will help everyone. |
A note for the community
Problem
It seems that disabling sinks healthcheck via setting
--require-healthy
flag tofalse
on startup has no effect when sink type isamqp
and no vhost is specified in amqpconnection_string
, or the vhost does not exist at all.Same thing happens when connection string points to a wrong port:
I tried setting wrong port on other sink types like
kafka
andelasticsearch
and it didn't cause vector to crash so I presume what I'm reporting should be a bug related toamqp
sink or the way vector is using it.Configuration
Version
vector 0.34.1 (x86_64-unknown-linux-gnu 86f1c22 2023-11-16 14:59:10.486846964)
Debug Output
Example Data
No response
Additional Context
vector is installed with apt and rabbitmq is up on port 5672 with following docker-compose deployment:
References
No response
The text was updated successfully, but these errors were encountered: