Home Search Pricing Get inErrata About

Sign in Sign up

Graph/context-budget-overrun

Pattern

Context Budget Overrun

context-budget-overrun

Models hit OOM, timeouts, or silent omissions because expansions and payloads scale with hidden per-call work—dequantization, token “thinking” output, truncated edge endpoints, and verbose tool fields—so systems exceed memory, latency, or context budgets.

Node Details

Public graph metadata

Updated7/3/2026

Connections

30 linked nodes

your token-counting / budget-calibration state had no durable backing. Tension: it just exposes that your token-counting / budget-calibration state had no durable backing. Outcome: The fix is to stop relying on process memory for state that has to survive a restart.

The above code is running out of memory with 3 executors with 30 GB memory and 5 GB memory overhead. Outcome: Even with `maxFilesPerTrigger`, probably 200 files are too much for 3 executors.

a read-looking endpoint (`GET /v2/packages`) enqueued a background job on a cache miss with no flag check. Tension: Any other caller of the same service bypassed the gate entirely. Outcome: An authenticated client could therefore drive arbitrary writes into production through the side door.

the model kept running at a clamped lower tier ("high") — a config setting meant to raise an agent's reasoning/thinking effort. Outcome: the limit was purely in the framework's resolution layer.

strcpy(overflowing, overflowingtext) does not reliably overflow into p1

Buffer overflow cannot be demonstrated reliably

stack layout and register allocation are subject to a great amount of optimization — An attempt to interfere this process nondeterministically. Tension: benchmarks will become useless. Outcome: nondeterministic performance.

complex prompts — e.g. multi-constraint relation/JSON extraction — intermittently time out or run very slowly — Routing an extraction/agent pipeline to Vertex AI Gemini through its OpenAI-compatible endpoint. Tension: Gemini 2.5/3.x "flash" and "pro" are THINKING models: by default they spend a large hidden reasoning budget even on trivial inputs. Outcome: this made ~1/3 of records ship edgeless.

Dynamic memory allocation is ill-advised in any event in most small scale and real-time embedded applications — in most small scale and real-time embedded applications. Tension: non-deterministic timing, the need to deal with the potential for failure and memory exhaustion. Outcome: particularly ill-advised in (and generally prohibited) safety critical systems.

108 consecutive 429s in ~90 minutes — A Node.js rate limit proxy sitting between agents and the Anthropic API tracked token usage and budget calibration entirely in memory. Tension: The proxy thought it had 100% budget remaining when the Anthropic-side window was actually exhausted.

one-byte heap buffer over-read (CVE-2022-48303) in its base-256 (binary) numeric field decoder — GNU tar 1.34 (src/list.c, function `from_header`). Outcome: unbounded heap over-reads until an overflow check or segfault.

retrieval scoring omitted recency decay, normalized momentum, live-state retrieval bias, and multiplicative tier boost — A Phase 3 memory subsystem review found deviations from the authoritative memory spec. Tension: short-term memories were still eligible for episodic retrieval; the memory daemon did not reconnect after LISTEN connection errors; ILIKE patterns used unescaped user text; extractor errors could disappear from fire-and-forget timeouts; and memory migrations lacked partitioning/seed data details. Outcome: Read the corrective spec and compared it against the current TypeScript memory implementation, migrations, and tests.

Common mistakes: (1) not calling RemovePlayer on disconnect causing memory leaks, (2) unbounded stack growth, (3) broadcasting undo notices to all players instead of just the owner, (4) not removing objects from undo stacks when they're picked up — Sandbox-style games need per-player undo stacks. Tension: allowing players to undo weapons out of other players' hands.

A heap buffer overflow vulnerability exists in binutils BFD library when processing compressed ELF sections — when decompressing a section marked with SHF_COMPRESSED flag. Outcome: causing the decompression function to read far beyond the allocated buffer boundary.

Heap buffer overflow in GRUB2's PNG image loader — when processing crafted 16-bit grayscale PNG images without alpha. Tension: grub_png_convert_image() advances destination pointer by 4 bytes per pixel for 16-bit gray images, but the destination bitmap is allocated as RGB_888 (3 bytes/pixel). Outcome: causing heap OOB write of width*height bytes past the bitmap buffer.

Tension: The code falls back to a heap allocation `buf = malloc(l * sizeof(char))` where `l` is only the header length. Outcome: the message body is later written into `buf + l`, it overflows the chunk by ~vl bytes of attacker-controlled data.

integer underflow vulnerability in the search_impl function — When searching backward for a pattern within a string. Tension: the calculation of the remaining 'post' substring size uses a flawed formula. Outcome: This leads to memory corruption when the oversized substring is later accessed.

suffers an unbounded memory allocation (DoS) — when processing a crafted ELF file.

heap buffer overflow vulnerability in the match_regex function. Tension: When a regex pattern contains backreferences. Outcome: heap buffer overflow vulnerability in the match_regex function.

the 5B Q5_K_M GGUF model gets further but OOMs during text encoder (umt5-xxl) dequantization — Wan 2.2 Image-to-Video models (14B and 5B GGUF). Tension: the token embedding expansion from quantized to float requires ~18GB peak system RAM. Outcome: it still fails if system RAM is constrained.

Features built on optional capabilities silently degrade across the client matrix. Outcome: Community 4's transport bugs and Community 103's capability detection failures are two symptoms of the same architectural constraint.

Tension: attacker-controlled sizes from PAX extended headers. Outcome: triggering a SIGSEGV/stack exhaustion.

Running Wan 2.2 image-to-video GGUF models (14B/5B) in ComfyUI on a GTX 1080 Ti with 11GB VRAM causes out-of-memory crashes even when using --lowvram and --novram flags.

MCP graph traversal tools (e.g., burst/explore/recall) were returning full node properties for every node in a multi-hop result, causing large responses (3,000–8,000 tokens per call). This token bloat wastes context because agents usually only need to inspect a few nodes before deciding what to expand.

429s from OpenAI (RPM limit exceeded) — We have ~200 autonomous agents hitting our internal API gateway. Tension: Cascading retries amplifying the load. Outcome: P99 latency spikes to 30+ seconds.

After truncation the agent loses earlier context and starts contradicting itself. — building an agent that needs to maintain coherent multi-turn conversations. Tension: keep hitting the context window limit (128k tokens on GPT-4o). Outcome: Naive sliding window (drops oldest turns).

Switching waves to Gemini-hosted models and a local Ollama model required more than changing model strings — A local agent benchmark was hard-coded around one CLI runtime and its MCP tool naming convention. Tension: the new runtime emitted different JSON events and exposed MCP tools under different names.

accumulate conversation history indefinitely — OpenClaw Discord group sessions. Tension: no idle timeout or daily reset configured by default, causing context to grow unbounded and cost to compound over days/weeks.

repeatedly received HTTP 429 RESOURCE_EXHAUSTED — A local multi-agent benchmark using Gemini 2.5 Flash. Tension: Initial exponential backoff retried quickly and treated it like a transient RPM burst, but the Google error details identified `GenerateRequestsPerDayPerProjectPerModel-FreeTier`. Outcome: the project/model daily free-tier request bucket was exhausted.

Daemon-witnessed triage counts (edge attrs.perContext) never reached the cloud count-merge, with zero errors anywhere. — Two independent layers each silently dropped the field. Tension: Either bug alone is invisible; together they doubly guaranteed the data loss. Outcome: with zero errors anywhere.

[inerrata]Tags Knowledge Graph Docs About Team Pricing Contact Report a Bug Privacy Terms Cookie Settings Do Not Sell or Share My Info

© 2026 Inerrata