fix(logs): sarama upgrade to 1.48.0 and panic recovery

[kongfei605]

1. 升级sarama到1.48.0 解决当前sarama版本bug的问题，sarama1.48.0要求go必须升级到1.25
2. 给日志采集pipeline相关的goroutine加上recover

[Copilot]

Pull request overview

This PR upgrades the Kafka client dependency (github.com/IBM/sarama) to v1.48.0 (and bumps the module’s Go version accordingly), and introduces panic recovery/restart wrappers for key goroutines in the logs pipeline. It also changes Kafka destination construction to return errors instead of panicking, so pipeline initialization can handle failures more gracefully.

Changes:

• Upgrade sarama to v1.48.0 and update related Go/module dependencies (including go 1.25.0).
• Add logs/util helpers (SafeGo, SafeGoWithRestart) to recover from panics and optionally restart goroutines with backoff, plus a Prometheus panic counter.
• Refactor Kafka destination creation to return (*Destination, error) and propagate/handle errors during pipeline startup.

Reviewed changes

Copilot reviewed 10 out of 11 changed files in this pull request and generated 3 comments.

Show a summary per file

File	Description
logs/util/recover.go	Adds SafeGo/SafeGoWithRestart wrappers with panic recovery and a Prometheus counter.
logs/sender/sender.go	Switches sender start to SafeGoWithRestart.
logs/processor/processor.go	Switches processor start to SafeGoWithRestart.
logs/pipeline/provider.go	Handles NewPipeline errors during pipeline initialization instead of assuming success.
logs/pipeline/pipeline.go	Updates NewPipeline to return error and propagates Kafka destination creation failures.
logs/input/file/tailer.go	Wraps tailer goroutines with SafeGo panic recovery.
logs/input/file/scanner.go	Switches scanner start to SafeGoWithRestart.
logs/client/kafka/producer.go	Runs async producer workers under SafeGoWithRestart.
logs/client/kafka/destination.go	Makes destination construction return errors; wraps async send goroutines with SafeGo.
go.mod	Bumps Go version and updates sarama / x/* dependencies.
go.sum	Updates sums for upgraded dependencies.

Comments suppressed due to low confidence (1)

logs/processor/processor.go:56

• SafeGoWithRestart will not successfully recover/restart if p.run panics because p.run's deferred p.done <- struct{}{} blocks until Stop is waiting. That can turn a panic into a permanent hang and also makes restarts unsafe (multiple run exits can block on done). Consider removing restart here or redesigning the done/stop signaling to be restart-safe (e.g., close-once semantics, buffered channel with sync.Once, or per-run done channels).

// Start starts the Processor.
func (p *Processor) Start() {
	util.SafeGoWithRestart("logs/processor", p.run, 5*time.Second)
}

// Stop stops the Processor,
// this call blocks until inputChan is flushed
func (p *Processor) Stop() {
	close(p.inputChan)
	<-p.done
}

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

Pull request overview

Copilot reviewed 11 out of 12 changed files in this pull request and generated 5 comments.

[Copilot]

Pull request overview

Copilot reviewed 12 out of 13 changed files in this pull request and generated 2 comments.

[Copilot]

Pull request overview

Copilot reviewed 12 out of 13 changed files in this pull request and generated 3 comments.

[Copilot]

Pull request overview

Copilot reviewed 12 out of 13 changed files in this pull request and generated 1 comment.

[Copilot]

Pull request overview

Copilot reviewed 12 out of 13 changed files in this pull request and generated 3 comments.