I have a Switchyard setup connected to a Bedrock Mantle endpoint, and even very simple requests such as "print hello" have been timing out or taking 30–60 seconds to complete throughout the day. Some requests never return at all. I'm testing the endpoint directly with curl. The regular, non-Mantle endpoint is responding normally. Is anyone else seeing similar throttling or elevated latency?
2 Answers
We’re seeing similar behavior and it doesn’t seem limited to a single request or prompt. Basic calls to the Mantle endpoint have been unusually slow or failing over the last couple of days, so this may be a broader service issue rather than a problem with your curl setup.
Our provisioned-capacity endpoint has continued responding at its normal speed. That suggests the problem may be isolated to the Mantle endpoint or a particular service path, rather than affecting Bedrock capacity generally.
That comparison is useful, although another endpoint in a different product path being healthy doesn’t necessarily rule out a Mantle-specific outage. The Mantle requests are still timing out for us.