Exchange - Which metrics are important to monitor / graph

Question

I have Exchange 2003 on a DC.

Here are the metrics which are available to me in order to make a graph out of them, or simply monitor their status:

EXNG: Inbound Connections Current
EXNG: Local Queue Length
EXNG: Local Retry Queue Length
EXNG: Memory Available Bytes
EXNG: Memory Pages\sec
EXNG: Messages Received/sec
EXNG: MSExchangeMTA Submits
EXNG: Page File Usage
EXNG: Pickup Directory Messages Retrieved/sec
EXNG: Remote Queue Length
EXNG: Remote Retry Queue Length
EXNG: Service State - Event
EXNG: Service State - IMAP4
EXNG: Service State - Information Store
EXNG: Service State - Management
EXNG: Service State - MTA Stacks
EXNG: Service State - Network News Transfer Protocol (NNTP)
EXNG: Service State - POP3
EXNG: Service State - Routing Engine
EXNG: Service State - Simple Mail Transfer Protocol (SMTP)
EXNG: Service State - Site Replication Service
EXNG: Service State - System Attendant

So which of these should I keep monitored and why ?

Answer 1

Service State - Information Store, Routing Engine, System Attendant, and SMTP will all tell you aggregately, generally, if the Exchange server is running/routing mail. If you have any POP3 or IMAP4 clients, you'd want to monitor that as well.

EXNG: Local Queue Length and Local Retry Queue Length will give you some information as to whether you're experiencing any mailflow issues inbound or outbound.

    EXNG: Memory Available Bytes
    EXNG: Memory Pages\sec
    EXNG: Messages Received/sec
    EXNG: Page File Usage

Should be monitored for general server health.