Conversation
|
| region (str): The AWS region to use for the AgentCore Runtime service. | ||
| session (Optional[boto3.Session]): Optional boto3 session. If not provided, | ||
| a new session will be created using default credentials. | ||
| region: AWS region name. If not provided, uses the session's region or "us-west-2". |
There was a problem hiding this comment.
I feel like we shouldn't default region, but I know this is against the existing pattern. @jariy17 what do you think?
| } | ||
|
|
||
| _ALLOWED_CP_METHODS = { | ||
| "create_agent_runtime", |
There was a problem hiding this comment.
q: how long does a runtime take to be ready? I know memory takes like 3 minutes so we have a create_and_wait utility. Wondering if the same could be useful here. Could come as a follow-up.
There was a problem hiding this comment.
Yep I had the same thought. Added utilities for wait on create
57614b8 to
22c09ed
Compare
tejaskash
left a comment
There was a problem hiding this comment.
Looks good. Follows the established MemoryClient passthrough pattern cleanly. Good test coverage across polling, timeouts, failure states, and teardown ordering.
Issue #: #391
Summary:
Previously, there was no way for customers to programmatically manage agent runtime resources (create, update, delete runtimes/endpoints) through the SDK. This PR adds passthrough plumbing, so allowlisted bedrock-agentcore-control and bedrock-agentcore methods are callable directly on AgentCoreRuntimeClient with snake_case kwarg support. This follows the same pattern established in MemoryClient.
Description of changes:
create_agent_runtimeupdate_agent_runtimeget_agent_runtimeget_agent_runtime_endpointdelete_agent_runtimedelete_agent_runtime_endpointinvoke_agent_runtimestop_runtime_sessionTest plan
By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.