High CPU load after upgrade to python client 1.2.0

[kmewhort]

Preflight checklist

• I could not find a solution in the existing issues, docs, nor discussions.
• I agree to follow this project's Code of Conduct.
• I have read and am following this repository's Contribution Guidelines.
• I have joined the Ory Community Slack.
• I am signed up to the Ory Security Patch Newsletter.

Ory Network Project

No response

Describe the bug

We're seeing major increases in CPU usage which we have traced to upgrading the python ory_kratos_client from 1.1.0 to 1.2.0. We are mainly just running ApiClient to_session calls (with lighter calls to the IdentityApi).

We have not yet done performance tracing to further narrow down the culprit between different Kratos calls, but just wondering if there's any obvious changes that could be culprit from 0.1.0 to 1.2.0?

Reproducing the bug

1. Run a set of calls using the following APIs, using the 1.1.0 client, while profiling CPU usage:

from ory_kratos_client.api import frontend_api, identity_api
from ory_kratos_client.exceptions import UnauthorizedException
from ory_kratos_client.models.create_identity_body import CreateIdentityBody
from ory_kratos_client.models.identity_with_credentials import IdentityWithCredentials
from ory_kratos_client.models.identity_with_credentials_password import (
    IdentityWithCredentialsPassword,
)
from ory_kratos_client.models.identity_with_credentials_password_config import (
    IdentityWithCredentialsPasswordConfig,
)
from ory_kratos_client.models.update_identity_body import UpdateIdentityBody

2. Run the same set of calls using the 1.2.0 client.
3. See that the CPU usage is much higher on the 1.2.0 client.

Relevant log output

Relevant configuration

Version

1.2.0

On which operating system are you observing this issue?

Linux

In which environment are you deploying?

Docker

Additional Context

No response

[aeneasr]

I don't think the SDK is at fault as it's just doing an HTTP call to our API. Maybe you can confirm that there are no other changes that may impact CPU on your end.

[kmewhort]

Thanks for the confirmation. Yes, looking at the lightweight library code, I concur it would be an unlikely culprit after all. We'll dive in elsewhere and I'll close this issue.