Setting TELEPORT_PROXY and TELEPORT_CLUSTER environment variables prevent tsh login --request-id from working

[Ezzahhh]

Expected behavior:
If TELEPORT_PROXY and TELEPORT_CLUSTER environment variables are set, it should still get updated certificates from the requested role. The behaviour should not differ from when the environment variables are not set (which works).

Current behavior:
If TELEPORT_PROXY and TELEPORT_CLUSTER environment variables are set, it is not possible to tsh login --request-id unless the variables are unset. If tsh request create is run and not cancelled or specified with --nowait then there are no issues pulling the right certificates after approval.

Bug details:

• Teleport version: 17.1.2

Recreation steps:

1. Do not set TELEPORT_PROXY and TELEPORT_CLUSTER environment variables
2. tsh login --proxy my-teleport-url:443 my-teleport-url
3. tsh request create --role my-role --reason='test'
4. Cancel out of the above with Ctrl+C
5. Approve the request with tctl
6. export TELEPORT_PROXY=my-teleport-url:443; export TELEPORT_CLUSTER=my-teleport-url;
7. tsh login --request-id=<request-id>
8. No log line Approval received, getting updated certificates...
9. Try to login to requested resource through role and get access denied

Audit Logs

{
  "cert_type": "user",
  "cluster_name": "teleport.{{REDACTED}}.com",
  "code": "TC000I",
  "ei": 0,
  "event": "cert.create",
  "identity": {
    "access_requests": []
// request-id is not contained here in this audit log when environment variables are set