Answer "how does this work?" from the code
Support gets the real answer from the product, not from an outdated doc.
Customer support · Software · When you ask · Updated
Works with
- Slack
- Notion
- Confluence
- Box
Slack
Notion
Confluence
BoxHow do you answer how-does-this-work questions from the code?
To answer how-does-this-work questions from the code, ask Maito in Slack. It reads the repository, finds the code that sets the behavior, and answers in plain words with the exact limit or rule. It also tells you when your docs disagree. You get an answer based on the product, not an old page.
How it's done today
Support asks engineering and waits.
What Maito does
- Read the question
- Check docs and the code
- Answer with what the product actually does
Example
@Maito does the API rate limit apply per user or per workspace? Pied Piper keeps getting 429s with only 3 users.
Per workspace: 600 requests a minute, shared by every user and token in it. That explains Pied Piper. Three users with scripts add up. The Notion page "API limits" is wrong. It says 100 requests a minute per user. The gateway code has no per-user counter at all. The limit is a single value in the rate limiter config, and the file hasn't changed since March. Also useful for the customer: every 429 includes a Retry-After header, so their scripts can back off instead of retrying at once. I read the code on main and the Notion page.
@Maito draft the fix for the page, I'll check with Leo.
Here is the replacement paragraph for the page. I haven't edited Notion. "Limits apply per workspace: 600 requests per minute across all users and tokens. Over the limit, the API returns 429 with a Retry-After header."
Sam (2:14 PM):
does the API rate limit apply per user or per workspace? Pied Piper keeps getting 429s with only 3 users.
Maito (2:17 PM):
Per workspace: 600 requests a minute, shared by every user and token in it. That explains Pied Piper. Three users with scripts add up.
The Notion page "API limits" is wrong. It says 100 requests a minute per user. The gateway code has no per-user counter at all. The limit is a single value in the rate limiter config, and the file hasn't changed since March.
Also useful for the customer: every 429 includes a Retry-After header, so their scripts can back off instead of retrying at once.
I read the code on main and the Notion page.
Sam (2:20 PM):
draft the fix for the page, I'll check with Leo.
Maito (2:21 PM):
Here is the replacement paragraph for the page. I haven't edited Notion.
"Limits apply per workspace: 600 requests per minute across all users and tokens. Over the limit, the API returns 429 with a Retry-After header."
How to set it up
- Set up
Nothing to set up. Ask in Slack or by email whenever you need it.
- Connect docs
- Notion
- Confluence
- Box
Not connected yet? Your agent sends you a link to sign in.
What it won't do
It reads the code on the branch you point it at. Behavior set outside the repository, such as a feature flag or a setting in a dashboard, is something it can't see.
Questions
They don't need to. Maito reads the code and explains the result without code snippets unless you ask.
It reads docs in Notion, Confluence or Box and says which pages disagree with the code.
No. It reads the code and docs only. Fixing a doc is a separate request.
