elastic / elastic/fleet-server
Fleet Server Native OTLP Output Support for Enhanced Gateway Integration
- Dominant language
- Go
- Stars
- 113
- Forks
- 117
- Avg merge
- 1d 16h
- Merged PRs (30d)
- 112
Description
This feature request proposes adding native OTLP (OpenTelemetry Protocol) output configuration support in Fleet Server to enable direct integration with EDOT Gateway deployments and external OpenTelemetry endpoints through both gRPC and HTTP protocols.
## Background
Currently, Fleet Server supports various output types for data forwarding, but lacks native OTLP output capabilities. While EDOT Collectors can be configured with OTLP outputs in standalone mode, this configuration cannot be managed centrally through Fleet Server's output settings.
The current Elastic documentation recommends [EDOT Gateway deployments](https://www.elastic.co/docs/reference/edot-collector/modes#gateway-requirements-for-self-managed-environments) for enterprise environments, and individual collectors can be configured with [OTLP outputs for gateway mode](https://www.elastic.co/docs/reference/edot-collector/config/default-config-standalone#gateway-mode). However, centralized management of these OTLP configurations through [Fleet Server output settings](https://www.elastic.co/docs/reference/fleet/fleet-settings#output-settings) is not currently available.
## Problem Statement
### Current Limitations
- **Configuration Fragmentation**: OTLP outputs must be configured individually per collector in standalone mode
- **Management Complexity**: No centralized control over OTLP endpoint configurations across managed agents
- **Deployment Inconsistency**: Different configuration approaches for gateway vs. direct output scenarios
- **Operational Overhead**: Manual configuration updates across multiple collectors when gateway endpoints change
### User Pain Points
- DevOps teams cannot centrally manage OTLP output configurations
- Difficult to implement consistent routing policies across all managed collectors
- Complex rollout procedures for gateway endpoint changes
- Limited flexibility in hybrid deployment scenarios (some agents to gateway, others to external endpoints)
## Proposed Solution
Add native OTLP output support to Fleet Server's output configuration, enabling centralized management of OpenTelemetry Protocol endpoints for all managed agents.
### Architecture Overview
```
┌─────────────────┐ OTLP ┌───────────────────┐
│ EDOT Collectors │ ────────► │ EDOT Gateway │
│ (Fleet Managed) │ │ (Fleet Managed) │
└─────────────────┘ └───────────────────┘
│
▼
┌───────────────────────┐
│ Backend Observability │
│ Cluster │
└───────────────────────┘
```
### Example
Contributor guide
Research direction
Start with Fleet Server's output settings and the linked Fleet output documentation, then trace how existing output types are configured and delivered. Define the gRPC and HTTP OTLP configuration requirements and confirm that managed agents can receive centralized gateway or external endpoint settings; done means both protocols are supported consistently through Fleet Server.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- go, grpc
- Domain
- backend, observability-sre
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Quiet
- Clarity
- Mostly clear
- Newbie friendliness
- 42/100