chore: test to prod

[mayankpande88]

No description provided.

[gemini-code-assist]

Summary of Changes

Hello @mayankpande88, I'm Gemini Code Assist¹! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed!

This pull request introduces a suite of enhancements focused on expanding observability capabilities, particularly for modern application stacks and cloud environments. Key updates include advanced eBPF-based L7 protocol detection, specialized monitoring for LLM APIs, and significant optimizations to Kubernetes integration and overall agent performance. These changes aim to provide more comprehensive and accurate insights into application behavior while improving the agent's efficiency and compatibility with newer infrastructure.

Highlights

• eBPF Tracing Enhancements: Significant improvements have been made to eBPF-based L7 protocol detection, including more robust handling for HTTP, Redis, PostgreSQL, RabbitMQ, ClickHouse, and the addition of FoundationDB. New monitoring capabilities for Node.js event loop blocked time and Python Global Interpreter Lock (GIL) contention have been introduced, alongside more reliable SSL/TLS library detection for tracing.
• LLM API Observability: The agent now provides dedicated metrics and tracing for Large Language Model (LLM) API calls, covering providers like OpenAI, Anthropic, Google, and Cohere. This includes tracking request counts, token usage (prompt, completion, total), and average latency, offering deeper insights into AI-driven applications.
• Kubernetes Integration & Resource Optimization: The Kubernetes IP resolver has been optimized by introducing a MinimalPod struct to drastically reduce memory consumption for cached Pod objects. Service-to-workload resolution has been enhanced, and the reliance on net.LookupAddr() has been removed in favor of more accurate eBPF-captured DNS data. Logging for verbose operations has been refined to use klog.V(5).Infof.
• Performance & Stability: Atomic throttling has been implemented for metric collection to prevent excessive resource usage during high-frequency scrapes. eBPF map processing has been optimized, and HTTP parsing now includes robust UTF-8 validation and improved header handling to prevent issues with malformed data. The /metrics endpoint now includes a 30-second timeout to enhance stability.
• Infrastructure & Compatibility Updates: The project's Go version has been updated to 1.24.7, and the Docker base image has been upgraded. The minimum supported Linux kernel version for eBPF functionality has been raised to 5.1. Support for Oracle Cloud Infrastructure (OCI) metadata discovery has also been added.

Ignored Files

• Ignored by pattern: .github/workflows/** (1)
  • .github/workflows/dev-ci.yaml

Using Gemini Code Assist

The full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips.

Invoking Gemini

You can request assistance from Gemini at any point by creating a comment using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands on the current page.

Feature	Command	Description
Code Review	/gemini review	Performs a code review for the current pull request in its current state.
Pull Request Summary	/gemini summary	Provides a summary of the current pull request in its current state.
Comment	@gemini-code-assist	Responds in comments when explicitly tagged, both in pull request comments and review comments.
Help	/gemini help	Displays a list of available commands.

Customization

To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a .gemini/ folder in the base of the repository. Detailed instructions can be found here.

Limitations & Feedback

Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for Github and other Google products, sign up here.

You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension.

Footnotes

1. Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩

[gemini-code-assist]

Code Review

This pull request introduces a substantial set of features and improvements, including support for FoundationDB, Node.js event loop monitoring, LLM API tracing, and Pressure Stall Information (PSI) metrics. It also includes significant performance and memory optimizations, such as using a minimal pod representation and an LRU cache for connection stats. The code is largely well-structured, but I've found a critical race condition in the metrics collection logic that needs to be addressed. I've also identified some areas with duplicated code and potential improvements for clarity and correctness.

[gemini-code-assist]

This function reads several fields of the Container struct (e.g., connectionStats, activeConnections, logParsers) without holding a lock. These fields are modified concurrently in event handlers (like onConnectionOpen, onProcessStart) which do hold a lock. This creates a race condition that can lead to panics or corrupted metrics. The c.lock.RLock() and defer c.lock.RUnlock() that were present in the previous version must be restored to ensure thread-safe access to shared data.

Suggested change

	c.lock.RLock()
	defer c.lock.RUnlock()

[gemini-code-assist]

The condition if len(parts) > 2 is confusing. parts is the result of strings.SplitN(line, ":", 3), and the code has already ensured len(parts) is 3, so this check is always true. It seems to be a typo and was likely intended to be if len(pp) > 1 to ensure the path has at least one component after splitting. This change would make the code clearer and more robust.

Suggested change

	if len(parts) > 2 {
	cgPath = "/" + pp[1]
	}
	if len(pp) > 1 {
	cgPath = "/" + pp[1]
	}

[gemini-code-assist]

This block of code for resolving a service to its underlying workload by matching pod labels is duplicated in handleServiceUpdate (lines 691-719). To improve maintainability and reduce redundancy, this logic should be extracted into a separate helper function, for example resolveServiceWorkload(service *v1.Service) Workload.

[gemini-code-assist]

The read_iovec function has been simplified to only process the first iovec entry. While this might be necessary to satisfy the eBPF verifier, it's a functional regression from iterating over MAX_IOVEC_SIZE entries. This limitation means that for writev or readv calls with multiple buffers, only the first buffer will be traced. It would be good to add a comment explaining this limitation and the reason for it.

// Only process the first iovec entry to avoid verifier issues with offset arithmetic.
// This is a limitation, as it means only the first buffer of a vectored I/O operation will be traced.
__u64 read_iovec(char *iovec, __u64 iovlen, __u64 ret, char *buf, __u64 *total_size) {

[gemini-code-assist]

Similar to read_iovec, this sys_enter_sendmmsg handler has been simplified to only process the first message header. This is a functional limitation for sendmmsg calls that send multiple messages at once. A comment explaining this limitation would be helpful for future maintainers.

// Only process the first mmsghdr to avoid verifier complexity.
// This is a limitation for sendmmsg calls with multiple messages.
int sys_enter_sendmmsg(struct trace_event_raw_sys_enter_rw__stub* ctx) {