System Monitoring Tool: MEDITECH Alert

MEDITECH Alert is a network monitoring system provided free of charge as part of the MIS portion of the Client/Server Release of MEDITECH's Health Care Information System. It provides warnings of problems for the components of a MEDITECH network including:

• File servers
• Application servers
• Background clients
• Network hardware components, including clients (PC or dumb terminal)
• MEDITECH applications.

When a problem condition occurs, or if a customer-defined threshold is met or exceeded, the system broadcasts an alert. Notification of the alert status is transmitted according to a series of user-defined parameters. MEDITECH Alert also provides a Daily Status Report which chronicles all alerts. The report runs automatically with its frequency defined by the user.

 

System Checks and Alerts

For file servers, the system conducts the following checks:

• File server alive and in the network being scanned
• VMagic Service live on the specified server
• Free space available on all file server drives including mirror drives
• File backups starts and stops
• VMagic database analysis starts and stops
• NT Log searches.

If any component in the system writes to the NT Log, the system can search for that event and can broadcast an alert based on criteria established by the user in the MEDITECH Alert File Server Dictionary.

For application servers, the system conducts the following checks:

• Application server alive and in the network being scanned
• VMagic Service is live and on the specified application server
• Free space on the C:\ drive of the application server.

For background clients, the system conducts the following checks:

• Background client is alive and in the network
• Daemon, or background manager, is alive and running
• Scheduled background job is running
• Jobs that fail to start as scheduled
• Transfers of background jobs to another client.

For networks, MEDITECH Alert conducts latency tests by performing pings on a group of representative clients. This determines how fast the network responds to client read requests. When network latency exceeds user-defined parameters, an alert is broadcast and the system continues to test the network until it identifies the cause(s), which may include:

• Intersegment delays caused by routers or bridges
• Slow file server response
• Decreased bandwidth availability.

The system also tests for the presence of clients on the network, identifying those that are down and not responding to network calls.

 

Application Checks and Alerts

The system conducts checks and provides more than 110 different alerts for all MEDITECH applications. User-defined application parameters are preset in the MEDITECH Alert dictionary. Some examples of situations that would prompt an alert include:

• Threshold of queues used by MEDITECH's Abstracting and Registration applications to communicate have been exceeded
• Month-end is out of balance in MEDITECH's Billing/Accounts Receivable application
• Oldest entry in a queue of a MEDITECH application has exceeded preset parameters.

The system takes advantage of its Microsoft^® Windows^® client operating system to provide an on-screen, graphical interface to all application database status screens. This interface gives users a single-screen view of all application database statuses and offers a quick, color-coded reference to the severity of the alert.

• Green: No alert, application operating within preset parameters
• Yellow: Low-level alert, one or more operations have reached or have slightly exceeded preset limits
• Red: High-level alert, one or more operations have exceeded preset limits

 

Standard Features

Users define who, how, and how frequently a notification action occurs by setting parameters in the system's Alert Notification dictionary. All of the system dictionaries link to the Alert Notification dictionary. The Alert Notification dictionary then determines the level of the alert and uses a combination of e-mail and paging to notify IS personnel on a defined hierarchical basis.

For some alerts (e.g., free disk space on a file server drive), varying levels of severity can be defined. Each level of severity carries with it a definition of who, how, and when to notify when an alert must be broadcast. Using free disk space on a file server partition as an example, one may see the following alerts broadcast:

• Low-level: 500 MB free, one e-mail message sent to main contact person
• Medium-level: 250 MB free, one e-mail message sent to several contact people and on-call person for MIS is paged one time
• High-level: Drive is almost out of free space, many contacts are sent an e-mail and several levels of on-call MIS personnel are paged every 15 minutes.

During an alert, MEDITECH Alert can invoke a program that shuts down the system to prevent others from signing on to it. By shutting down the system automatically, there is no chance for corrupt data since no additional data can be entered. This spares MIS from having to conduct checks for data integrity.

There are several types of alerts that do not have varying levels (e.g., VMagic service not running on an application server) because there are only two states for the component being monitored -- on and off. In those cases, the system issues the highest level of alert defined for that component.

A daily status report offers a high-level look at all alerts and notifications. The report provides a summary of everything that happened in the system since the last report was run. The reports run automatically according to the frequency defined by the user in the MEDITECH Alert dictionary.