RPC Retry & Timeout Support

[adhdtech]

Add retry and timeout options to DRP_Endpoint.SendCmd. Now that DRP_CmdError is used as a standard error class, return codes can be evaluated to determine whether or not a retry should be attempted.

A retry can be initiated by any of these conditions:

• Timeout expiring (execution of ServiceCmd starts timer)
• Error Codes: 408, 500, 503, 504, 507

Error codes other than those listed should result in a normal response to the client. If the retry threshold is reached, the last error code will be returned to the caller.

Regarding errors kicked back locally:

• An RPC call timeout should result in a 408.
• An error sending the command to the endpoint (missing, etc) should result in a 500.

Decisions to make:

• Should the timeout parameter be passed on to downstream calls?
• Should the timeout parameter be added as a new attribute or as part of DRP_Cmd.params?

[adhdtech]

Update the ReplyHandlerQueue structure. Something along these lines.

Existing:
{ $token: $cb }

New:
{ $token: { sendTime: $timestamp, timeout: $timeout, callback: $cb, type: "rpc" | "stream" } }