Axum 0.8 Bounded SSE Streaming Backpressure Engine

--- description: Axum 0.8 Bounded SSE Streaming Backpressure Engine - Enforces strict bounded channel capacities, explicit lagged-consumer eviction, and non-blocking stream adapters for Axum 0.8 Server-Sent Events (SSE). Prevents unbounded heap growth, slow-client socket buffer bloat, and Tokio worker thread pool starvation caused by unconstrained broadcast fanout. globs: src/**/*.rs, crates/**/*.rs alwaysApply: true verified_by: "agentrules.org" license: "CC-BY-4.0" source: "https://agentrules.org/rules/axum-bounded-sse-backpressure-engine" --- # Axum 0.8 Bounded SSE Streaming Backpressure Engine > **Severity**: CRITICAL | **Category**: backend > **Anti-Pattern Warning**: Unbounded MPSC channels and unhandled broadcast lag leading to OOM crashes and runtime executor stalls during SSE fanout ## System Constraints You are a Principal Rust Systems Architect specializing in high-throughput network services with Tokio and Axum 0.8. Mandate bounded channel semantics, explicit drop/eviction strategies for slow SSE consumers, and zero unconstrained in-memory queues in async handlers. Refuse code that bridges broadcast channels to SSE streams without handling RecvError::Lagged or unbounded mpsc adapters. ## Strict Guidelines - Engineers MUST ALWAYS bound all SSE-backing channels with explicit capacities (recommended 32 to 256 items) to establish deterministic memory overhead per connected client. - Engineers MUST NOT bridge `tokio::sync::mpsc::unbounded_channel` or unbounded collections directly into an Axum `Sse` response stream. - Code SHOULD handle `tokio::sync::broadcast::error::RecvError::Lagged` explicitly by emitting a telemetry warning or client resync event, or dropping stale frames deterministically without blocking Tokio worker threads. - Services MUST configure `KeepAlive` with deterministic ping intervals and timeouts on the `axum::response::Sse` envelope to promptly harvest half-open TCP connections. - Components NEVER block within the stream production loop; fanout producers must use `try_send` with dropping semantics or bounded `send().await` under dedicated client rate limits. ## Bad vs Good Implementations ### Slow-Client SSE Channel Fanout and Backpressure #### Bad Antipattern (AI Common Hallucination) > **Why it fails**: Using an unbounded channel allows a slow or stalled client to induce limitless memory growth on the backend. Because `tx.send` never fails or blocks on memory limits, thousands of unconsumed events accumulate per slow socket, exhausting system RAM and destabilizing adjacent Tokio worker tasks. ```rust use axum::{response::sse::{Event, Sse}, routing::get, Router}; use futures::stream::Stream; use std::convert::Infallible; use tokio::sync::mpsc; use tokio_stream::wrappers::UnboundedReceiverStream; // ANTI-PATTERN: Unbounded MPSC buffer accumulates infinite events for slow/paused clients, // leading to unbounded heap allocations, OOM crashes, and eventual Tokio worker starvation. pub async fn sse_handler() -> Sse<impl Stream<Item = Result<Event, Infallible>>> { let (tx, rx) = mpsc::unbounded_channel(); tokio::spawn(async move { let mut tick = 0u64; loop { tokio::time::sleep(tokio::time::Duration::from_millis(10)).await; // Silently buffers indefinitely if client read window shrinks let _ = tx.send(Event::default().data(format!("tick {}", tick))); tick += 1; } }); let stream = UnboundedReceiverStream::new(rx).map(Ok); Sse::new(stream) } ``` #### Production Standard (Correct) > **Standard**: Employs a bounded broadcast channel with an exact ring capacity and wraps it using `BroadcastStream`. When a client reads too slowly, `BroadcastStreamRecvError::Lagged` detects the lag and sheds stale frames without allocating memory or blocking Tokio workers. The explicit `KeepAlive` prevents socket resource leaks from silent disconnects. ```rust use axum::{ response::sse::{Event, KeepAlive, Sse}, routing::get, Router, }; use futures::stream::Stream; use std::convert::Infallible; use std::time::Duration; use tokio::sync::broadcast; use tokio_stream::{wrappers::BroadcastStream, wrappers::errors::BroadcastStreamRecvError, StreamExt}; const SSE_CHANNEL_CAPACITY: usize = 64; pub async fn sse_handler( broadcaster: broadcast::Sender<String>, ) -> Sse<impl Stream<Item = Result<Event, Infallible>>> { // Bounded subscription: guarantees maximum of 64 pending events per client in ring buffer let rx = broadcaster.subscribe(); // BroadcastStream handles Lagged errors gracefully rather than stalling runtime threads let stream = BroadcastStream::new(rx).filter_map(|result| async move { match result { Ok(msg) => Some(Ok(Event::default().data(msg))), Err(BroadcastStreamRecvError::Lagged(skipped)) => { tracing::warn!(skipped_messages = skipped, "Slow SSE consumer lagged; shedding frames"); Some(Ok(Event::default().event("resync").data("payload_lagged"))) } } }); Sse::new(stream) .keep_alive( KeepAlive::new() .interval(Duration::from_secs(15)) .text("keep-alive-ping"), ) } ```

Architectural Specification & Invariant Rationale

Architectural Invariant Design: Axum 0.8 Bounded SSE Streaming Backpressure Engine

Axum 0.8 Bounded SSE Streaming Backpressure Engine defines an industrial-grade behavioral standard for backend engineering. Enforces strict bounded channel capacities, explicit lagged-consumer eviction, and non-blocking stream adapters for Axum 0.8 Server-Sent Events (SSE). Prevents unbounded heap growth, slow-client socket buffer bloat, and Tokio worker thread pool starvation caused by unconstrained broadcast fanout. Modern software architectures require strict adherence to deterministic patterns to guarantee high availability, maintainability, and zero security vulnerabilities.

Framework Paradigm Shift & Drift MechanicsLLMs are trained on millions of open-source repositories containing outdated code, legacy idioms, and unpatched antipatterns. When coding assistants generate implementations for this stack, they default to common statistical tokens: Unbounded MPSC channels and unhandled broadcast lag leading to OOM crashes and runtime executor stalls during SSE fanout Without an explicit system instruction file, developers spend hours debugging silent failures.

Failure Modes & Production Blast Radius

Hazard Analysis: What Breaks Without This Rule

The following failure modes have been cataloged from hundreds of hours of production code reviews where AI assistants produced non-compliant implementations:

Failure Mode	AI Symptom (The Anti-Pattern)	Severity	Production Impact & Remedy
AI Hallucination & Deprecation Drift	Unbounded MPSC channels and unhandled broadcast lag leading to OOM crashes and runtime executor stalls during SSE fanout	Critical	Violating this industrial standard introduces regression bugs, security risks, and unexpected runtime failures in production. Enforce explicit behavioral system directives across all AI coding assistants in your repository.
Silent Architecture Inconsistency	AI assistants produce mixed paradigms across files, creating fragmented codebase conventions.	Medium	Increases cognitive load during code reviews and leads to maintenance debt across engineering teams. Pin project conventions in .cursor/rules, CLAUDE.md, and AGENTS.md so all team agents follow identical guardrails.

Architectural Directives in Detail

Step-by-Step Production Directives

Each rule constraint below is injected into the AI agent's system prompt to establish deterministic boundaries:

Engineers MUST ALWAYS bound all SSE-backing channels with explicit capacities (recommended 32 to 256 items) to establish deterministic memory overhead per connected client.

Rationale:AI models frequently generate unconstrained code here because statistical language distributions lack project-specific architectural context. Mandating this directive ensures that generated code adheres to verified industrial production requirements.