Make logs consistent

[flavio]

Currently the logs produced by the agent and by the controller are not consistent.

Some log stament produced by the agent are json encoded, while others are not.

Agent logs - example-1

Mixed bag of json and text:

time=2026-02-05T13:46:00.857Z level=INFO msg="Probing eBPF features..." component=agent
time=2026-02-05T13:46:01.192Z level=INFO msg="Cgroupv2 supported controllers detected successfully" component=agent cgroup.controllers="[cpuset cpu io memory hugetlb pids rdma misc dmem]"
time=2026-02-05T13:46:01.192Z level=INFO msg="Load time configuration detected" component=agent component=ebpf-manager cgrp_fs_magic=cgroupv2 cgrp_v1_subsys_idx=0 debug_mode=0
time=2026-02-05T13:46:01.480Z level=INFO msg="learning mode is enabled" component=agent
time=2026-02-05T13:46:01.481Z level=INFO msg="connecting to NRI socket" component=agent component=nri-handler
time=2026-02-05T13:46:01.533Z level=INFO msg="starting manager" component=agent
{"level":"info","ts":"2026-02-05T13:46:01Z","msg":"starting server","name":"health probe","addr":"[::]:8081"}
{"level":"info","ts":"2026-02-05T13:46:01Z","logger":"controller-runtime.metrics","msg":"Starting metrics server"}
{"level":"info","ts":"2026-02-05T13:46:01Z","logger":"controller-runtime.metrics","msg":"Serving metrics server","bindAddress":":8080","secure":false}
time=2026-02-05T13:46:01.636Z level=INFO msg="Starting BPF Manager..." component=agent component=ebpf-manager
{"level":"info","ts":"2026-02-05T13:46:01Z","msg":"Starting EventSource","controller":"learningEvent","source":"channel source: 0xc000186230"}
{"level":"info","ts":"2026-02-05T13:46:01Z","msg":"Starting Controller","controller":"learningEvent"}
{"level":"info","ts":"2026-02-05T13:46:01Z","msg":"Starting workers","controller":"learningEvent","worker count":1}
time="2026-02-05T13:46:01Z" level=info msg="Created plugin 00-agent (agent, handles StartContainer,RemoveContainer)"
time=2026-02-05T13:46:01.642Z level=INFO msg="Starting gRPC exporter" component=agent component=grpc_exporter addr=:50051 mTLS=true

Agent logs - example-2

Some log levels are uppercase, others are not:

time=2026-02-05T13:46:01.642Z level=INFO msg="Starting gRPC exporter" component=agent component=grpc_exporter addr=:50051 mTLS=true
time="2026-02-05T13:46:01Z" level=info msg="Registering plugin 00-agent..."

The same seems to happen also inside of the controller.

Controller logs

2026-02-05T13:45:28Z	INFO	controller-runtime.webhook	Serving webhook server	{"host": "", "port": 9443}
2026-02-05T13:45:28Z	INFO	controller-runtime.certwatcher	Starting certificate poll+watcher	{"cert": "/tmp/k8s-webhook-server/serving-certs/tls.crt", "key": "/tmp/k8s-webhook-server/serving-certs/tls.key", "interval": "10s"}

I0205 13:45:28.853226      43 leaderelection.go:258] "Attempting to acquire leader lease..." lock="runtime-enforcer/4e873589.rancher.io"

2026-02-05T13:45:29Z	INFO	controller-runtime.metrics	Serving metrics server	{"bindAddress": ":8443", "secure": true}

I0205 13:45:46.670555      43 leaderelection.go:272] "Successfully acquired lease" lock="runtime-enforcer/4e873589.rancher.io"

2026-02-05T13:45:46Z	INFO	WorkloadPolicyStatusSync	Starting with	{"interval": "30s"}

Acceptance Criteria

• Use structured logging everwhere
• Use JSON format by default
• Use Go's log/slog, there are adapters also for controller runtime (see what sbomscanner is doing)