What is the issue?

Users have reported WARNs being emitted in application logs for trust_dns warnings. We can see messages like the following in the application logs:

[ 656.276898s] DEBUG ThreadId(01) trust_dns_proto::tcp::tcp_stream: zero bytes read, stream closed?
[ 656.276905s] WARN ThreadId(01) trust_dns_proto::xfer::dns_multiplexer: stream TCP(10.100.0.10:53) error: stream closed

We should add trust_dns=ERROR to our default log filter to exclude these types of messages and avoid confusion until the issue is resolved upstream (tracked here)

How can it be reproduced?

Dump destination controller logs and look for trust_dns messages whenever a query is sent.

Logs, error output, etc

[ 656.276905s] WARN ThreadId(01) trust_dns_proto::xfer::dns_multiplexer: stream TCP(10.100.0.10:53) error: stream closed

output of linkerd check -o short

n/a

Environment

n/a

Possible solution

Add trust_dns=ERROR to our default log filter to exclude these types of messages until issue is properly resolved upstream.

Additional context

No response

Would you like to work on fixing this bug?

None