TLS handshake timeout: transport timeouts too aggressive; requests fail in CI

[bwalsh]

Summary

Requests to calypr-dev.ohsu.edu are failing with a net/http: TLS handshake timeout error when running in CI/CD environments. The current transport-level timeouts may be too aggressive for the network, especially given variability in CI network performance and hospital/institutional infrastructure.

Error details

Observed error:

level=ERROR source=/home/runner/go/pkg/mod/github.com/hashicorp/go-retryablehttp@v0.7.8/client.go:735 msg="request failed" pid=801632 error="Post \"https://calypr-dev.ohsu.edu/user/credentials/api/access_token\": net/http: TLS handshake timeout" method=POST url=https://calypr-dev.ohsu.edu/user/credentials/api/access_token

Current transport config (request/request.go)

DialContext Timeout:      5s
TLSHandshakeTimeout:     5s
ResponseHeaderTimeout:   10s

Proposed fix

Increase transport timeouts to allow for higher-latency environments:

• Dialer.Timeout: 10s
• TLSHandshakeTimeout: 15s
• ResponseHeaderTimeout: 30s

This should prevent premature TLS handshake failures and better accommodate slow network situations in CI and institutional networks.

Additional context

• The http.Client global Timeout is set to 0 (no upper bound), so transport timeouts are critical.
• Retries will not help if each handshake fails due to a short timeout; all will fail the same way.

---

Request: Update request/request.go to use more permissive transport-level timeouts for reliability.

[bwalsh]