Mongo timeouts to Atlas from tsh db connect

[stevenGravy]

Expected behavior:

A method of increasing the timeout to a MongoDB.

Current behavior:

Users that connect via tsh db connect experience timeouts when connecting to Atlas. This is more pronounced in Teleport Cloud even with AWS Global Balancing enabled. The tsh db connect to a MongoDB atlas DB can experience timeouts since it defaults to 2000 ms in the mongo shell invocation. Note that the direction

Bug details:

• Teleport version 9.3.3
• Recreation steps

Open a MongoDB Atlas account. Connect a Teleport cloud instance to it. tsh db connects will connect w/o timing out half the time.

• Debug logs

$ tsh db connect --db-user=mongouser  mongodb-atlas
Current Mongosh Log ID:	6298cf9b669c3e9914acf0d9
Connecting to:		mongodb://localhost:51230/?directConnection=true&serverSelectionTimeoutMS=2000&tls=true&tlsCertificateKeyFile=%2FUsers%2Fstevenmartinsr%2F.tsh%2Fkeys%2Fskyward.teleport.sh%2Fsteven-db%2Fskyward.teleport.sh%2Fmongodb-atlas-x509.pem&tlsCAFile=%2FUsers%2Fstevenmartinsr%2F.tsh%2Fkeys%2Fskyward.teleport.sh%2Fcas%2Fskyward.teleport.sh.pem&appName=mongosh+1.4.2
MongoServerSelectionError: Server selection timed out after 2000 ms
ERROR: exit status 1