Persistenza errore "Reconnecting: stream disconnected before completion: response.failed event received"
Resolved 💬 4 comments Opened Feb 11, 2026 by digifra Closed Feb 12, 2026
💡 Likely answer: A maintainer (github-actions[bot], contributor)
responded on this thread — see the highlighted reply below.
allego l'id del feedback con i log della sessione 019c4cfd-f48d-7742-8f76-3b21daf8000e
4 Comments
Potential duplicates detected. Please review them and close your issue if it is a duplicate.
Powered by Codex Action
Please share any relevant parts of your config.toml file (redacting any personal information). It looks like you're attempting to use a custom model? Are you using an Azure endpoint?
Hi,
Here are the relevant parts of my config.toml (with sensitive data omitted):
----
model = "gpt-5.1-codex-max"
model_reasoning_effort = "medium"
model_provider = "azure"
tool_output_token_limit = 4096
[model_providers.azure]
name = "azure"
base_url = "https://nome-risorsa.openai.azure.com/openai"
env_key = "AZURE_OPENAI_API_KEY"
query_params = { api-version = "2025-04-01-preview" }
wire_api = "responses"
----
Yes, I'm using a custom model configuration and an Azure OpenAI endpoint. The same error happens also with gpt-5.2-codex.
Let me know without problems if you need attitional informations.
Thanks in advance.
Ah, you're using an Azure-hosted model. We've received similar reports of this problem on Azure here. Please report the issue to Azure to make sure they're aware of it.