Bug Description

Dear Hermes Integration Team,

I am reporting a recurring availability issue when using Hermes Agent with MiniMax M2.7. Every API call to the endpoint https://api.minimax.io/anthropic fails with:

⚠️ API call failed (attempt 1/3): OverloadedError [HTTP 529]
The server cluster is currently under high load. Please retry after a short wait ...

This happens repeatedly, even with moderate usage and short prompts, and appears to be a service‑sided overload condition rather than a user‑rate‑limit violation. The retries in Hermes simply keep hitting the same 529, which severely degrades usability.

Could you please:

Investigate whether this is a capacity or routing issue on the MiniMax side when invoked via Hermes, and

Clarify if this is expected behavior under current load, or if there are configuration changes (e.g., endpoint, timeout, or request‑format) we should apply on our side?

If you have a preferred channel for API‑availability reports (GitHub issue, Discord, or support email), kindly let me know so I can route similar incidents there in the future.

Best regards,
Kenneth

Location: Hong Kong
Tool: Hermes Agent (Terminal)
MiniMax model: MiniMax‑M2.7
Approximate occurrence time: April 15, 2026, around 5:15 PM HKT

Steps to Reproduce

Dear Hermes Integration Team,

I am reporting a recurring availability issue when using Hermes Agent with MiniMax M2.7. Every API call to the endpoint https://api.minimax.io/anthropic fails with:

⚠️ API call failed (attempt 1/3): OverloadedError [HTTP 529]
The server cluster is currently under high load. Please retry after a short wait ...

This happens repeatedly, even with moderate usage and short prompts, and appears to be a service‑sided overload condition rather than a user‑rate‑limit violation. The retries in Hermes simply keep hitting the same 529, which severely degrades usability.

Could you please:

Investigate whether this is a capacity or routing issue on the MiniMax side when invoked via Hermes, and

Clarify if this is expected behavior under current load, or if there are configuration changes (e.g., endpoint, timeout, or request‑format) we should apply on our side?

If you have a preferred channel for API‑availability reports (GitHub issue, Discord, or support email), kindly let me know so I can route similar incidents there in the future.

Best regards,
Kenneth

Location: Hong Kong
Tool: Hermes Agent (Terminal)
MiniMax model: MiniMax‑M2.7
Approximate occurrence time: April 15, 2026, around 5:15 PM HKT

Expected Behavior

no disconnection!

Actual Behavior

Dear Hermes Integration Team,

I am reporting a recurring availability issue when using Hermes Agent with MiniMax M2.7. Every API call to the endpoint https://api.minimax.io/anthropic fails with:

⚠️ API call failed (attempt 1/3): OverloadedError [HTTP 529]
The server cluster is currently under high load. Please retry after a short wait ...

This happens repeatedly, even with moderate usage and short prompts, and appears to be a service‑sided overload condition rather than a user‑rate‑limit violation. The retries in Hermes simply keep hitting the same 529, which severely degrades usability.

Could you please:

Investigate whether this is a capacity or routing issue on the MiniMax side when invoked via Hermes, and

Clarify if this is expected behavior under current load, or if there are configuration changes (e.g., endpoint, timeout, or request‑format) we should apply on our side?

If you have a preferred channel for API‑availability reports (GitHub issue, Discord, or support email), kindly let me know so I can route similar incidents there in the future.

Best regards,
Kenneth

Location: Hong Kong
Tool: Hermes Agent (Terminal)
MiniMax model: MiniMax‑M2.7
Approximate occurrence time: April 15, 2026, around 5:15 PM HKT

Affected Component

CLI (interactive chat)

Messaging Platform (if gateway-related)

No response

Debug Report

Dear Hermes Integration Team,

I am reporting a recurring availability issue when using Hermes Agent with MiniMax M2.7. Every API call to the endpoint https://api.minimax.io/anthropic fails with:

⚠️ API call failed (attempt 1/3): OverloadedError [HTTP 529]
The server cluster is currently under high load. Please retry after a short wait ...

This happens repeatedly, even with moderate usage and short prompts, and appears to be a service‑sided overload condition rather than a user‑rate‑limit violation. The retries in Hermes simply keep hitting the same 529, which severely degrades usability.

Could you please:

Investigate whether this is a capacity or routing issue on the MiniMax side when invoked via Hermes, and

Clarify if this is expected behavior under current load, or if there are configuration changes (e.g., endpoint, timeout, or request‑format) we should apply on our side?

If you have a preferred channel for API‑availability reports (GitHub issue, Discord, or support email), kindly let me know so I can route similar incidents there in the future.

Best regards,
Kenneth


Location: Hong Kong
Tool: Hermes Agent (Terminal)
MiniMax model: MiniMax‑M2.7
Approximate occurrence time: April 15, 2026, around 5:15 PM HKT

Operating System

Mac26.3.1

Python Version

No response

Hermes Version

No response

Additional Logs / Traceback (optional)

Root Cause Analysis (optional)

No response

Proposed Fix (optional)

No response

Are you willing to submit a PR for this?

• I'd like to fix this myself and submit a PR