When I visit and/or refresh the log list, my latest logs are often not showing… and sometimes, even after applying a filter that indicates there should be results, the results are not being rendered.
A picture speaks 1000 words:
1 Like
I’m “authorized”. I’ve made a buncha Responses calls in the last few hours to document being overbilled on “detail”:“low” images, and overlooked setting “store”:false or capturing the ID.
— Testing gpt-5.1-codex (Responses)
Input tokens: 370
The image is a black-and-white checkerboard pattern composed of equal-sized squares arranged in alternating colors. It resembles a classic chess or checker board, extending across the entire frame in a grid of contrasting black and white tiles.— Testing gpt-5.1-codex-mini (Responses)
Input tokens: 344
The image appears to be entirely blank white with no visible objects or details.— Testing gpt-5.1-2025-11-13 (Responses)
Input tokens: 370
The image shows a simple black‑and‑white checkerboard pattern: evenly spaced square tiles alternating black and white in a grid that fills the entire image.— Testing gpt-5.1-chat-latest (Responses)
Input tokens: 370
It’s a black‑and‑white checkerboard pattern made up of evenly sized squares arranged in a grid. The squares alternate colors in both the horizontal and vertical directions.
Nothing new showing up:
Delays again after “we fixed it” in other topics.
The background error is not the culprit, just bad programming. The code of the platform site always is hitting a bunch of endpoints that have nothing to do with logs or even “API reference”. I get 401 also, because of Bearer null header being sent:
message: "Incorrect API key provided: null. You can find your API key at https://platform.openai.com/account/api-keys.", type: "invalid_request_error", code: "invalid_api_key", … }
That endpoint is specifically for listing legacy user “profile” keys here:
https://platform.openai.com/settings/profile/api-keys
And is hit multiple times with the correct session key when in the right place.
1 Like
Hi and welcome to the community!
I can partly reproduce what you’re seeing. When I view the logs without filters, everything appears after a short delay. But once I filter by the specific model, the expected results are indicated yet not actually shown.
running into the same issue for at least 2 days and its really frustrating.
i dont normally add any filters, and the logs dashboard always shows stale rows, almost like some sort of bad cache.
i’ve had little luck getting around this, very few times I tried changing different filters and date ranges which seemed to reset the cache and get me fresh entries. but it’s so annoying as I test and debug things and constatnly need to check fresh logs
please help fix
Hey everyone, Apologies for the inconvenience. We tried reproducing the issue on our end but we can see the logs for response api now. Can someone please confirm if this issue is resolved on your end as well. Please make sure to clear cache and try reviewing the logs.
seems to be working for me now, thanks for the quick response!
The logs site still has issues. It takes adding a date range, which is unsuccessful, then removing a date range, or other such manipulations to force more queries to finally receive a listing by chance that is somewhat recent.
Thus, any log display (Responses) remains untrustworthy.
sigh, seems like the dashboard is showing stale rows again…
edit: I’m getting fresh logs every few refreshes. Is it possible we’re being routed across prod servers with inconsistent state?
I am still not getting desired results… the logs list is not reliably showing me the most current logs. See attached:
apologies for the 3 posts… the system wouldn’t let me embed multiple images into my one reply because I’m a “new user”… how insulting!
1 Like
Hey thanks for sharing the screenshots!
It’s still kind of glitched on my end as well but differently.
Let’s hope this gets fixed soon!
Regarding the post restrictions: the system works very well for moderating a large online community. It’s not perfect though.
1 Like
Hey everyone, We are not able to reproduce this issue on our end. The logs are appearing more reliably now. Please confirm, if that is not the case. Thank you!
1 Like
Conclusion, working good on Responses
- Open new browser tab with platform.openai.com; no other tabs for the site;
- Wait for the spinning “signing in” for cookies to turn into auth and session token
- click “dashboard”
- click “logs”
- “Completions” is the default view, with some nov 19 fine-tuning runs (unwanted, from alternating endpoints on the chat playground)
- twenty-four Dec 3 results are included, which weren’t on December 3.
- a December-only date range query also includes them, viewed today on December 8.
- Comment out “store” in the same script, test API 9:11am, finish another round of calls 9:14am (exactly 5 days later!)
- Write up report, open dev tools for network logging, and clear the prior date range,
9:16AM: Two API calls are made, with query parameters that don’t make obvious sense:
https://api.openai.com/v1/responses?include[]=message.input_image.image_url&input_item_limit=1&output_item_limit=1
https://api.openai.com/v1/responses?after=resp_0ba918511513ec5d00693706e3c8888197afd16706afd31228&include[]=message.input_image.image_url&input_item_limit=1&output_item_limit=1
- See the logged Responses API calls.
Good Job!
Issue:
Two codex models fail with Status:500, an internal error on a API request, where successes was received before:
— Testing gpt-5.1-codex-max (Responses)
gpt-5.1-codex-max Image usage: 353 Image prompt tokens: 353 Total Usage: 360— Testing gpt-5.1-codex (Responses)
— ERROR: gpt-5.1-codex: Error code: 500 - {‘error’: {‘message’: ‘An error occurred while processing your request. You can retry your request, or contact us through our help center at help.openai.com if the error persists. Please include the request ID req_d4e681ad878746908ff1d60c9f32da9b in your message.’, ‘type’: ‘server_error’, ‘param’: None, ‘code’: ‘server_error’}}— Testing gpt-5.1-codex-mini (Responses)
— ERROR: gpt-5.1-codex-mini: Error code: 500 - {‘error’: {‘message’: ‘An error occurred while processing your request. You can retry your request, or contact us through our help center at help.openai.com if the error persists. Please include the request ID req_9a67105c53674cf69d86f053da74bb9b in your message.’, ‘type’: ‘server_error’, ‘param’: None, ‘code’: ‘server_error’}}
This seems to have been a passing fluke, as more attempts are successful, and appear with no obvious delay when again clearing the logs date range for a refresh:
1 Like
The log system is lately working much better for me. Logs of new requests seem to appear upon refresh almost immediately.
One improvement suggestion, an inline “refresh” icon in the list view, instead of having to refresh the whole browser just to retrieve the latest logs.
1 Like
I also can no longer reproduce the issue on my end. The problem was a similar error message after a filter had been removed, but it now works as expected.
Amazing, Thanks for confirming guys. I will close this thread for now.
1 Like