I am getting a rate limit error on the Chart Generation lab of Module 2:
RateLimitError: Error code: 429 - {'error': {'message': 'exceeded quota for this month'}}
I have only run the scripts of the jupyter notebook once, so this doesn’t make sense. My account is 2 days old. Can someone help?
Hi @rasty_penny welcome to the deeplearning.ai community.
The 429 - exceeded quota for this month error is a platform-side issue, meaning your code and your 2-day-old account are completely fine.
The DeepLearning.AI Jupyter labs route requests through shared, built-in API keys provided by the course. When you see this specific error, it means the platform’s shared key for the Chart Generation lab has temporarily hit its monthly budget due to total learner traffic.
Next steps:
Wait 24–48 hours: I am escalating this to the DeepLearning.AI Community Team so the engineering staff can provision more API quota. It typically takes a day or two for them to refresh the backend keys.
Do not modify your code: Your current setup is correct. Just leave the notebook as-is.
Do not use a personal key: There is no need to create a personal OpenAI account, add billing details, or paste your own API key into the environment.
Take a break for now, and try re-running the lab tomorrow. Reply back here if it is still blocking you after a couple of days!
@Community-Team @chris.favila the keys for Module 2 of this course needs a refresh, can somebody please help ?
Hi Sanjay, thanks for the response.
1) Regarding M2_UGL2:
I am now able to execute the code, but I am getting a different error:
TypeError: datetime64 type does not support sum operations
it seems that the LLM is ignoring instruction 8:
Any advise?
2) Regarding C1M2_Assignement:
When running the unit test, I am getting once again the RateLimitError:
Got: An error occurred: Error code: 429 - {‘error’: {‘message’: ‘exceeded quota for this month’}}
This is a very serious issue. How are we meant to finish the course if we all share a monthly budget and we have no visibility at all on what is happening in the background?
I’m also seeing this error, any expectation on how long will the keys take to renew?
Thanks!
i am not ready to wait 24-48 hrs. any better solution than one at my cost?
Hi @rasty_penny — on the datetime64 does not support sum operations error: it’s not actually the LLM ignoring instruction 8 (the string-concatenation warning). The model is correctly avoiding date + time concatenation, but it’s making a separate mistake in the aggregation step:
grouped_sales = q1_sales.groupby(['year', 'coffee_name']).sum().reset_index()
.sum() here is called on the whole grouped DataFrame instead of just the price column. Since date is still a datetime64 column at that point, pandas can’t sum it and throws the TypeError — nothing to do with concatenation.
Quick fix, works without waiting on a course update — select the column first:
grouped_sales = q1_sales.groupby(['year', 'coffee_name'])['price'].sum().reset_index()
If you can’t edit the LLM-generated cell directly, you can patch it after extraction instead:
match = re.search(r"<execute_python>([\s\S]*?)</execute_python>", code_v1)
if match:
initial_code = match.group(1).strip()
initial_code = initial_code.replace(
"grouped_sales = q1_sales.groupby(['year', 'coffee_name']).sum().reset_index()",
"grouped_sales = q1_sales.groupby(['year', 'coffee_name'])['price'].sum().reset_index()"
)
exec_globals = {"df": df}
exec(initial_code, exec_globals)
(If your generated code phrases the line slightly differently, adjust the .replace() strings to match exactly.)
Unrelated to this: I also posted the same root-cause note over in the M2_UGL_1 thread if you want more context — that one’s specific to the datetime bug, this thread’s mostly about the quota issue.
Hi. Tagging @lesly.zerna here to have a look at the quota. Thanks for reporting.
@lesly.zerna
FYI: There are multiple reports of rate limiting in the Agentic AI course, since three days ago.
Oh I see. Will help escalate this to more people. Thanks for reporting!
Hi everyone. Please try again now. Thank you.
Hi Chris.
I tried again. Everything was working fine to begin with. I completed the 1st ungraded lab of module 3 without issues, but when I proceeded to the 2nd ungraded lab of module 3, I encountered the same error:
LLMError: An error occurred: Error code: 429 - {‘error’: {‘message’: ‘exceeded quota for this month’}}
Now I cannot proceed to the graded labs!!
Copy. Thank you for the details.
Hi @lesly.zerna . Looks like there’s still an issue in the lab mentioned here.
Just to clarify: the problem is not limited to a specific lab. Once I encounter the exceeded quota error in one lab, then I encounter the same error in all other labs I attempt to run!
I tried to complete the first lab today and ran into the same two issues.
In Step 2, the initially generated code failed with the error “TypeError: datetime64 type does not support sum operations.” The generated aggregation tried to sum all remaining columns, including the date column.
Later, when I tried to run the complete workflow, the first model call failed with “RateLimitError: Error code: 429 — exceeded quota for this month.” As a result, I couldn’t complete the end-to-end workflow because the lab environment had already exhausted its monthly quota. And this is just the very begining of the course.
@lesly.zerna What I’m supposed to do? Just wait?
Thanks for reporting!
Hi everyone! the fixed for this Ungraded Lab in Module 2 have been pushed to prod so you can see changes reflected in the notebook and currently working as expected!
Thank you for your patience!
-- Lesly, DLAI
I’m getting this error each time I run an inference request in the labs. Currently I’m in M5 Agentic AI - Market Research Team.
LLMError: An error occurred: Error code: 429 - {‘error’: {‘message’: ‘exceeded quota for this month’}}
Hi Nicolas!
Let me report it to our engineering team. Thanks for letting me know.
Just researching… have you been trying running all notebooks from this Agentic Course today , notebooks worked until you got to M5? or you just (re)started from M5 and got the error?
I’ve seen this in other modules and the issue started last week during another module and has persisted since.
Unfortunately I’m stuck on this lab with the same error. Do I need to wait until November?? @lesly.zerna
Well, that might be the solution by now. 