Resolved
This incident has been resolved.
Monitoring
the issue causing latency across various models has been resolved.
Monitoring
performance has normalized across qwen/qwen3-32b, openai/gpt-20b, openai/gpt-120b and llama-3.1-8B-instance. The team is continuing to monitor.
Monitoring
We are seeing latency issues starting to recover across openai/gpt-oss-20b, llama-3.1-8B-instance, and qwen/qwen3-32b following a fix that has been implemented by the team. We continue to monitor to ensure the issue is fully remediated.
Investigating
The current models experiencing latency issues are: openai/gpt-oss-20b, llama-3.1-8B-instance, and qwen/qwen3-32b. They team is continuing to implement a fix.
Investigating
We are currently investigating latency across various models. TTFT may be impacted. Our engineers are analyzing current system performance and dependencies to isolate the issue.