Uptime monitoring services

A third-party uptime monitoring service monitors service providers' infrastructures, alerting them when failures are detected and maintaining a record of uptime performance.

Such a service has two primary uses:

To alert service providers when a failure has occurred so that they can take corrective action. It is vital that the relevant systems administrator receives failure notifications quickly, so SMS text messages tend to be an important channel through which alerts are sent.

To monitor and report on the ongoing uptime performance of service providers. Both providers and their customers can then use this information to evaluate whether SLAs (Service Level Agreements) and uptime guarantees have been met.

[edit]

What should be monitored?

Different service providers have different monitoring needs. However, service providers tend to offer a general set of monitoring tests, including (but not limited to):

HTTP Request: tests for a correct HTTP response from a web server.
SMTP: tests the ability of an SMTP server to accept email messages.
POP3: tests the ability of a POP3 server to accept connections and client logins.
IMAP: tests the ability of an IMAP server to accept connections and client logins.
FTP: tests the correct functionality of an FTP server to accept connections and authenticate.
DNS: tests that a DNS server will response correctly to domain name lookup requests.
TCP: a general TCP connection test, allowing arbitrary services to be tested.

The methodology to determine what constitutes a "failure" varies wildly, so it is crucial that the monitoring provider communicates precisely how failures are detected and confirmed before alerts are sent and downtime is recorded.

It is generally accepted that failures should be confirmed in multiple, geographically disparate locations before alerts are sent. Hence most uptime monitoring providers use multiple points of presence to facilitate this.

[edit]