People ask us this all the time, and while we put our new website together, we thought it would be a good idea to breakdown all the things that Sentinel can keep tabs on for you. So grab a coffee & stay focused because it’s a big old list.
With Sentinel you can monitor a wide variety of services both external (HTTP, POP3, SMTP etc) and internal to your server (DISK, MEMORY etc).
All of these services can be monitored at time intervals right up to every minute and alerts can be delivered via any combination of SMS, Email, Twitter DM, iPhone Push (via Prowl) and our API.
Every service that you monitor with Sentinel have base settings that include;
- Friendly name – A way to give the service a more meaningful name. Very useful if you have multiple instances of a service on the same server.
- Confirmation checks – The maximum number of times for Sentinel to check a server or service to confirm the problem exists, before sending out an alert.
- Confirmation retry interval – The number of minutes to wait between confirmation checks.
- Reminder interval – The duration to wait after we send the first alert, before reminding you again about the problem.
- Flap Detection – The formula we use to determine if a service is rapidly changing between OK and non-OK states. Can be disabled it not required.
Below is a comprehensive list of all the services that the Sentinel Server and Website Monitoring system can currently monitor.
External services
These are services that are exposed to the Internet. Each have default options and some have additional parameters that you can specify too. These services are able to be monitored on any server operating system.
PING
PING is the default service that we add every time you add a server to monitor on Sentinel (by default Sentinel will PING your server every 5 minutes). The PING check can be disabled at anytime as long as there is another check type in place for that server.
Options for the PING check:
- Warning settings for Latency & Packet loss – When this threshold is reached Sentinel will send out a WARNING alert.
- Critical settings for Latency & Packet Loss – When this threshold is reached Sentinel will send out a CRITICAL alert.
- Packets – Number of PING packets to send.
HTTP
HTTP is the standard check we use to monitor a website. It has quite a few options and good to know that you can have MULTIPLE HTTP checks per server if you want to. This is very good if you have multiple pages (or HTTP servers) on your server that you wish to monitor.
Options for the HTTP check:
- Uri – The web page that you want to check. The default setting is “/” but you can check any page that you like.. eg /shopping-cart/ or /sign-up.php. This is quite useful to ensure that critical web pages on your site are actually there and haven’t been removed or replaced with a generic ‘Your site has been hacked’ page.
- Vhost – The (virtual) host name you wish to check (default is the server name). You can monitor as many virtual hosts as you like on each server.
- Port – The network port number to check (default is port 80).
- Warning Latency – The number of seconds it takes to get a response before sending a warning alert. Disabled by default.
- Critical Latency – The number of seconds it takes to get a response before sending a critical alert. Disabled by default.
HTTPS
HTTPS is the standard check we use to monitor secure HTTP websites.
HTTPS options include:
- Port – Set to 443 by default but you can change that if you like.
- Uri – Works the same way as the Uri for the HTTP check.
SMTP
SMTP is the check that we use to make sure that your incoming mail server is available.
The only option currently configurable for SMTP is the Port (set at 25 by default).
IMAP
If you run an IMAP server or just need to check that mail is available for your customers then this is the check for you.
The options you can change on this are:
- Port – Set to 143 by default.
- Username – If you specify a username and a password then Sentinel can test that your IMAP server is actually authenticating accounts.
- Encryption – You can tell Sentinel whether or not to use encryption to connect to your IMAP server. Options here are, None (default), SSL and TLS.
POP3
If you run a POP3 server or just need to check that mail is available for your customers then this is the check for you. The options are the same as the IMAP server above in that you can specify the following…
- Port – Set to 110 by default.
- Username – If you specify a username and a password then Sentinel can test that your IMAP server is actually authenticating accounts.
- Encryption – You can tell Sentinel whether or not to use encryption to connect to your IMAP server. Options here are, None (default), SSL and TLS.
CERTIFICATE
CERTIFICATE is an interesting check but surprisingly handy. What it does is simply remind you BEFORE your SSL certificates expire. Very handy indeed! The certificate check frequency is set to ‘Every day’ by default (rather than ‘Every 5 minutes’ for all the other services). This check allows you to check for HTTPS, POP3 and IMAP secure certificate expiry to ensure your clients applications don’t get met with nasty error messages!
Options include:
- Type – Choose either HTTPS, POP3S, or IMAPS.
- Warning – The number of days before expiry to send out a WARNING alert.
- Critical – The number of days before expiry to send out a CRITICAL alert.
DNS
With Sentinel you can check that the DNS service is running and able to return the DNS entry for a given domain name.
Options include:
- Host – The domain name to check for on your DNS server.
- Timeout – The number of seconds to wait before the check times out. (10 by default).
FTP
Sentinel can check that your FTP service is available. At the moment we don’t support authentication checking for FTP and the only setting that you can change is the Port (21 by default).
SSH
Sentinel can monitor Secure Shell to check that it is available on your server.
Options include:
- Port – The default port is 22.
STREAM (Icecast & SHOUTCast for monitoring streaming radio)
Developed from a customer request, this is perfect if you run an Internet radio station or other streaming audio system. You can monitor MULTIPLE STREAMS on the one server.
Options include:
- Type – Either Icecast or SHOUTCast
- Uri – The stream location. By default this is “/”
- Port – The port that the stream is running over. By default this is 80.
TCP (generic service)
Sentinel allows you to monitor the transport layer over any port you like.
Services that run on your Server
By installing a small piece of software on your servers (the Sentinel AGENT software) you can then use Sentinel to monitor and alert you about physical properties and services running on your servers.
The Sentinel AGENT runs on most current Linux Distributions, Mac OS X 10.5+ and Windows operating systems. Below is a table of the services that we can monitor on each of these operating systems.
| Linux | Mac OS X | Windows | |
| DISK SPACE | X | X | X |
| LOAD | X | X | X |
| PROCESSES | X | X | X |
| SWAP SPACE | X | X | |
| MEMORY | X | ||
| MAIL QUEUE | X | X | |
| NTP | X | X |
DISK SPACE
Sentinel can check the available disk space on any partition on your server and it can send you a WARNING alert when you have a certain percentage (or megabytes) of free disk space remaining and a CRITICAL alert when you have a certain percentage of free disk space remaining.
Options include:
- Partition (or Disk) to check – By default this is “/”
- Warning – The warning %, or megabyte amount of free space remaining (Default is 20%)
- Critical – The critical %, or megabyte amount of free space remaining (default is 10%)
LOAD
Sentinel can check the LOAD on your servers and alert you when WARNING and CRITICAL thresholds have been reached.
Options include:
- Warning Threshold – The warning load level for 1min, 5min and 15min averages
- Critical Threshold – The critical load level for 1min, 5min and 15 min averages
PROCESSES
PROCS is very useful to ensure that certain processes are running on your server. A classic example would be to use the PROCS check to make sure MySQL or CRON was running on your server.
Options include:
- Process – The name of the process on the server. eg: myslqd
- Range – The range of processes to be considered OK (defaults to 1:65535) eg: 1:10 would alert you if there is LESS than 1 process running or MORE then 10 processes running
SWAP SPACE
Sentinel can tell you if your swap space drops below a certain percentage (indicating that memory is getting dumped to swap). The default percentages are to send a WARNING alert if swap drops below 90% and a CRITICAL alert if swap drops below 80%.
MEMORY
Sentinel can monitor MEMORY usage on Windows servers and alert you when certain thresholds are reached.
MAIL QUEUE
Sentinel can check the amount of Mail in the queue for various common Mail applications and you can specify WARNING and CRITICAL thresholds to send alerts.
Options include:
- MTA – Mail Transfer Agent can be either Postfix, Sendmail, Exim or QMAIL.
- Warning – Warn if there are more than x number of messages in the mail queue.
- Critical – Warn if there are more than x number of messages in the mail queue.
NTP – Network Time Protocol
We can check your server clock against a NTP host server to ensure that your system clock is nice and accurate.
Options include:
- Hostname – The host name of the NTP server you wish to check your clock against.
- Warning – The number of offset seconds that should be considered a warning (defaults to 0.5).
- Critical – The number of offset seconds that should be considered critical (defaults to 1).
{ 8 comments… read them below or add one }
Looks really good, might have to trial it and see how I can best incorp it into our current structure. -Thanks
Hey Cass,
Thanks for the feedback. Give it a trial and see how it goes.
There is a bunch of stuff we are doing (and launching soon) that will be ideal for business like yours.
Follow us on twitter ( @sentinelmonitor ) and we’ll keep you in the loop and feel free to email me directly – mark@sentinelmonitoring.com – if you have any questions.
I’m not seeing any description of what is actually monitored on a web URL, is it simply checking for a non-error response header? I need the ability to check for a keyword in the data/page response to see if i’m getting the response required, i have services or pages that may return an error response to the web browser/user.
Hi Scott, At the moment we don’t have the option to check for a keyword in the data/page response with our standard HTTP check.
You could probably use a CUSTOM check for that though. There is a bit more information on that here. http://sentinelmonitoring.com/faq/checks/can-you-monitor-custom-services/article/39
Yes this would be possible for what i’m thinking. how does your software determine if a page is replaced with ‘you have been hacked’ example mentioned in your docs? just wondering what web site monitor checks are being done other than no response/time out, latency, header error status. thanks!
Hi Scot,
In the example above we’re talking about being able to check for a specific page on your site. Perhaps the example isn’t very clear.
What happened to me once was that I installed a wordpress site and came back to it a couple of days later to find that someone had hacked the site and that every attempt to get to a page like /about or /contact simply wouldn’t work – just redirected to a generic page that said ‘You’ve been hacked’.
The site still responded fine so Sentinel didn’t raise any alarms. So we extended Sentinel to check for the existence of specific pages on a site – and if they couldn’t be found – then an alarm would be raised
Is there any possibility to write custom scripts and run them by Sentinel? To make custom parameters checks. Similar as it could be configured for the Nagios for example.
Hi Maksym,
We do have something we call a ‘CUSTOM’ check. That will essentially allow you to write your own scripts and return results which Sentinel can monitor. Check out this FAQ Article and if you have any questions please drop us an email to: support@sentinelmonitoring.com
http://sentinelmonitoring.com/faq/checks/can-you-monitor-custom-services/article/39