Managed Monitoring (Server Monitoring)

FAQs

• Why do I need Managed Monitoring?
• Can't I just purchase or build my own monitoring server?
• If my server or a service running on my server fails, how long will it take to receive the notification?
• What types of alerts will I receive?
• Will I be notified when the server/service recovers?
• What if the server/service is down for an extended period of time?
• Can I temporarily disable a service check or a notification?
• What types of devices/services can you monitor?
• I'm running a special server - can you monitor it for me?
• Can I use this service to monitor systems located anywhere on the Internet?
• My servers are protected by a firewall which will block your monitoring server. What should I do?
• What if your monitor server fails?
• Do you provide historical statistics on the availability of each of my services being monitored?
• How do I maintain my monitoring notifications and alerts?
• What types of notifications do you support?
• How many notifications can I define per service check?
• How many service checks can I setup?
• How often is the "state" information on the portal updated? How often is the historical uptime data updated?
• What are the timeout settings for the various service checks which must be met before a service is considered to be down?
• I have multiple employees, each of whom are on call different days of the week and different times of day. Can I configure the system so each person only gets notified of outages during their on-call periods?
• Can I define "maintenance windows" so I don't get alerted to failures while doing maintenance on my servers?

Why do I need Managed Monitoring?

If you have one or more servers or devices connected to the Internet and maximum availability of those devices is important to you, you need to ensure your systems are monitored 24 hours per day, 7 days per week. Network and server outages are a fact of life. Causes include power outages, hardware failure, human error, hacker attacks, and software problems.

If you rely on your servers for providing mission critical services or even if you simply run a small e-commerce operation, server downtime will cost you money. Once you understand that outages can and will happen, you need to put systems in place to ensure that outages are detected quickly and the problems are resolved quickly.

Primus' "Managed Monitoring" service can monitor all of your systems and alert you when failures occur.

Can't I just purchase or build my own monitoring server?

Absolutely. However, the cost and expense to purchase, install, configure, and maintain the hardware and software to monitor your network infrastructure would be significantly more expensive than the low monthly fee for Primus' managed monitoring service. The cost for a dedicated phone line alone to support paging or faxing would be almost more than the cost of Primus' service. Primus' monitoring server is located on Primus' high-capacity, redundant network in our ultra-secure and reliable Internet Data Center.

If my server or a service running on my server fails, how long will it take to receive the notification.

The timing of the notifications depend on a number of items:

1. How often your service is being monitoring (every 1 minute or every 5 minutes).
2. The timeout configured on the particular monitor which has to be exceeded before the system will consider the service "failed".
3. Whether the notification is being sent by email, fax, or pager.
4. The time required for you to receive the email, fax, or pager after it has been sent. For example, it may take two to three minutes to receive a fax. Some paging networks may take several minutes to deliver a page once they receive it. If your email server is slow, the delivery of the email alert may be delayed.

What types of alerts will I receive?

You will receive "Critical" alerts when the service you are monitoring fails. When the service recovers, you will receive an "Ok" alert. You may also receive a "Warning" alert for certain services if they appear to be failing.

Will I be notified when the server/service recovers?

Once your system recovers, the monitoring server will send you an "OK" message indicating that your system is once again responding normally.

What if the server/service is down for an extended period of time?

Primus' monitoring server will re-send "CRITICAL" alerts every 6 hours if your system is down for many hours or days.

Can I temporarily disable a service check or a notification?

Yes, you can. You might want to disable a particular notification for a period of time if the person who normally receives the notification is going to be away on vacation or for some other reason. You might want to disable a particular service check for a period of time if you are going to be doing maintenance on the server or taking it out of production for a period of time.

What types of devices/services can you monitor?

Primus' monitoring server supports all popular Internet protocols. These include ping (icmp), http, http-ssl, ftp, dns, smtp, ntp, imap, pop, real media, ldap, postgresql, mysql, oracle, generic tcp, http content and http-ssl content. You can specify a custom port number for any of these checks. The http content and http-ssl content checks allow you to specify a particular webpage URL and a particular text string which the page should contain. If the page does not contain the specified text string, the service will be considered "down".

In addition to the above protocols, the service can also monitor servers and network devices using SNMP. The system can query SNMP variables on your device and alert if the returned values do not match predetermined values or ranges.

The service can also monitor a vast number of metrics for UNIX and Windows servers by way of a plugin which you install on your server(s). Using this feature, you can monitor items such as disk space consumption, CPU utilization, whether a process or service is running, system uptime, etc.

I'm running a special server. Can you monitor it for me?

Quite possibly. Contact Primus' Corporate Sales Team for more details.

Note that you can use the generic TCP check to at least tell whether or not a service is responding on a particular port.

If you have a more complicated server environment, you can use the http content check to monitor deeper into your network. For example, if you have one or more web servers which rely on a back-end database server, you could write a simple CGI script and install it on your web servers to verify that your database server is responding properly. If it is, the CGI would always output a text string like "success" which the http content monitor would check for. If your CGI detects a problem with the database server, it could output "error" and the content monitor would consider the service to be "down" and generate an alert.

Can I use this service to monitor systems located anywhere on the Internet?

Yes. Primus does not restrict the monitoring to devices located just on Primus' network. You can monitor systems located elsewhere on the Internet.

My servers are protected by a firewall which will block your monitoring server. What should I do?

Primus can provide you with the IP address of the monitoring server. You can add a rule to your firewall to allow the particular traffic from that IP address through the firewall required for the monitor service to perform the checks.

What if your monitor server fails?

Primus' monitor server features redundant components and is itself monitored by additional monitoring systems 24x7. The server is located in Primus' Class "A" Internet Data Center. This facility features redundant power, redundant Internet connections, and a temperature and humidity-controlled high security environment.

If Primus does need to perform maintenance on the monitoring server or a problem occurs with the server, Primus will notify you by email to alert you to the outage and to provide an estimated time to recovery.

Do you provide historical statistics on the "availability" of each of my services being monitored?

For each service being monitored, you are provided with two pieces of information which are updated daily - total days/hours/minutes "up" and total days/hours/minutes "down".

How do I maintain my monitoring notifications and alerts?

You will use Primus' customer portal to define all of your notifications and alerts. You can also access recent "state" information for each alert via the portal as well as historical "uptime" data.

What types of notifications do you support?

You can define email, fax, and pager notifications. Primus currently supports sending alphanumeric pages through the Bell Mobility and Rogers Wireless paging networks.

How many notifications can I define per service check?

You can define as many notifications as you wish per service check. That is, if you want to have an email sent to multiple email addresses or if you want to have multiple pages or faxes sent when a service fails, you can do so.

How many service checks can I setup?

The quantity of service checks you can define is limited by the package level you purchase from Primus.

How often is the "state" information on the portal updated? How often is the historical uptime data updated?

A snapshot of the "state" information is pushed from the monitoring server to the portal every 10 minutes. This information is not meant to be used as a real-time view of the health of your service checks.

The historical uptime data is updated every 24 hours.

What are the timeout settings for the various service checks which must be met before a service is considered to be down?

¹ The NTP check compares the time received from the NTP server vs. the time on Primus' monitoring server. If the time is off by more than 60 seconds, a "Warning" alert will be generated. If the time is off by more than 120 seconds, a "Critical" alert will be generated.

I have multiple employees, each of whom are on call different days of the week and different times of day. Can I configure the system so each person only gets notified of outages during their on-call periods?

Yes. You can define custom "time periods" which are essentially "monitoring windows" during which alerting will occur. For each custom "time period", you can define "monitoring windows" for the seven days of the week. A "monitoring window" can span the entire day (for each day of the week), part of each day, or no time period of a particular day.

Once you define a custom time period, you can associate it with a "notification" or a "service check". This gives you total flexibility when setting up your monitoring entries.

Can I define "maintenance windows" so I don't get alerted to failures while doing maintenance on my servers?

You have two options for maintenance windows. You could either log in and flag a service check as "inactive" before you start your maintenance window, or if the maintenance window is a re-occurring event, you could define a custom time period which would prevent alerts from being generated during the time period you specify.