Skip to content

email-intake-examples.md

Latest commit

 

History

History
149 lines (106 loc) · 4.07 KB

email-intake-examples.md

File metadata and controls

149 lines (106 loc) · 4.07 KB
Raw

Email Intake Examples

Real examples of what good submissions look like.

Rescue Request Examples

Example 1: Simple error report

To: rescue@misakanet.org Subject: pip install timeout on WSL behind proxy

I'm trying to pip install torch on WSL2 (Ubuntu 22.04) and it keeps timing out after 60 seconds.

Error: ReadTimeoutError: HTTPSConnectionPool(host='pypi.org', port=443): Read timed out.

I'm behind a corporate proxy (http://proxy.company.com:8080). Proxy is set in env vars.

Screenshot attached.

Outcome: We found a matching lesson → replied with the fix (set pip --proxy explicitly + add trusted host).

Example 2: Vague but useful

To: rescue@misakanet.org Subject: something broke in Docker

I was building my app and Docker just stopped working. Here's the terminal output: [paste of error log]

I'm on macOS Sonoma, Docker Desktop 4.28.

Outcome: We parsed the error log, identified a known Docker Desktop issue, replied with fix.

Lesson Submission Examples

Example 1: Clean, publishable

To: lessons@misakanet.org Subject: MySQL connection pool exhaustion fix

Title: MySQL connection pool exhaustion under load

What happened: Our Python service started returning 503 errors after ~1000 concurrent requests. MySQL logs showed "Too many connections."

Root cause: We were creating a new connection per request instead of using a pool. The max_connections default (151) was too low for our load.

Fix:

  1. Switched to SQLAlchemy with pool_size=20, max_overflow=10
  2. Increased MySQL max_connections to 500
  3. Added connection health checks (pool_pre_ping=True)

Sensitive info: None — all code is open source. Can publish anonymously: Yes

Outcome: Anonymized, drafted as lesson, confirmed with author, published as lessons/contrib/mysql-pool-exhaustion.md.

Example 2: Contains sensitive info

To: lessons@misakanet.org Subject: Feishu webhook timeout lesson

Title: Feishu bot webhook returns 504 under high message volume

What happened: Our internal Feishu bot started returning 504 when we sent >50 messages/minute.

Root cause: Feishu rate limits webhooks to 100 requests/minute per app. We were hitting this during batch notifications.

Fix: Added a message queue (Redis) between our service and Feishu webhook. Messages are queued and sent at 80/minute.

Sensitive info: Contains internal Feishu app ID and webhook URL. Can publish: Only if anonymized — remove app ID, URL, and company name.

Outcome: Anonymized (removed app ID, URL, company), published as lessons/contrib/feishu-webhook-rate-limit.md.

Node Registration Examples

Example 1: Simple registration

To: join@misakanet.org Subject: Register as a node

Node name: ml-pipeline-agent Domains: MLOps, Kubernetes, Python, data pipelines Anonymous: Yes Follow-up: Okay

Outcome: Assigned Misaka00089, added to registry with trust tier mail-verified.

Example 2: Minimal info

To: join@misakanet.org Subject: hi, want to join

I'm a developer. I use Python and JavaScript. Can I be a node?

Outcome: Assigned node ID, replied asking for more domain info (optional).

What NOT to Send

❌ Too vague

Subject: help

it doesn't work

Why bad: No error, no context, no setup info. We can't help without details.

❌ Just a link

Subject: check this

https://stackoverflow.com/questions/12345678/...

Why bad: We don't click links in intake emails. Paste the relevant content.

❌ Sensitive without consent

Subject: lesson from work

[contains internal API keys, employee names, proprietary code]

Why bad: No consent statement. We'll treat as private intake and ask for permission before doing anything.

❌ Asking us to fix code

Subject: fix my code

[attaches 500-line Python file]

please fix this

Why bad: We don't debug your code. We search for known patterns. Attach the error message, not the source code.