Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!


Uptime Monitor Feature Requests?
New on LowEndTalk? Please Register and read our Community Rules.

All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.

Uptime Monitor Feature Requests?

Hi all,

Looking for some features requests for an uptime monitor. I have a project in progress right now that I'm not 100% ready to release, but suffice to say it's an uptime and performance monitor, so to speak. I'm looking for features that users of LET might be interested in, and as a bonus anyone who responds will be given a lifetime account upon request when launched.

Just to note: this isn't just another vapourware or useless project. It has been meticulously thought through and engineered and any false downtime reports once launched will be refunded in the form of a free beer (which I set at $3 paid by PayPal).

(Launch will occur in beta form shortly and will be announced on LET, and full launch in the new year.)

--Adam

Thanked by 1lukesUbuntu
«1

Comments

  • Please consider adding thresholds for latency, i.e. Ping takes greater than xy.z milliseconds, or Http get takes longer than x.yz seconds.

    Also, It would be great to be able to fetch a file and parse values from that file.
    i.e. http get http://server.tld/status.txt and regexp parse the contents:
    ^Loadavg: [1-9].[0-9]+$ (would trigger if load average > .9 )

    Thanked by 1rmlhhd
  • It would also be great if it could monitor services. E.g. Apache/Nginx on or off.

    Also network speed, how much traffic is currently going in/out.

  • Ping requests and TCP/UDP requests notification when over x milliseconds? Done. Can't promise the same for DNS/other requests. Also, HTTP get file and parse results: Done. That'll be there on launch.

    If you're interested when launched, you're good for an unlimited lifetime account. You can message me on launch or if you want me to message you drop me a message now (I don't want to bother anyone without permission).

    -- Adam

    Thanked by 1jimpop
  • @rmlhhd Do you mean port monitoring of port 80 (for example)? Otherwise do you mean connect via SSH and check services up (doable, but different).

    Network speed, also doable, but you'd need to have a test-file available. Is that OK? For example: you could set a HTTP check to get http://www.example.com/1GB.test and it'd record the speed of the transfer.

    --Adam

  • @adly said:
    rmlhhd Do you mean port monitoring of port 80 (for example)? Otherwise do you mean connect via SSH and check services up (doable, but different).

    Network speed, also doable, but you'd need to have a test-file available. Is that OK? For example: you could set a HTTP check to get http://www.example.com/1GB.test and it'd record the speed of the transfer.

    --Adam

    Yeah I mean check services up but I suppose if your checking Apache or MySQL then you could check ports 80 & 3306

    Network usage is what I mean, bandwidth used or current network load.

  • @rmlhhd Network usage is an interesting one. I considered an option to install a node service that monitored how things were running, but figured nobody wanted it. More than happy to offer such an option of people want it.

    If I'm misunderstanding, please let me know.

    --Adam

  • matthewvzmatthewvz Member, Host Rep

    Game server monitoring would be amazing. I would love a central place to see all of my player counts and uptime.

  • Maybe something to make sure port 25 is working for SMTP. Something simple to replace a lot of functions nagios checks

  • Maybe go as far as do a SSH login and run a command and check the result as a monitor? That way a webserver doesn't need to be setup?

  • @pcfreak30 said:
    Maybe go as far as do a SSH login and run a command and check the result as a monitor?

    I like this idea (please make sure you support ssh keys). Also, the ability to take action and execute commands via an ssh session. For instance, if latency is high to a server, I would like the remote ability to ssh in and shut it down or restart, etc.

  • @pcfreak30 said:
    Maybe go as far as do a SSH login and run a command and check the result as a monitor? That way a webserver doesn't need to be setup?

    That's an interesting idea. I implemented connecting to SSH and checking for a prompt, but not running a command. Is this something people are really interested in? Would a custom script be useful to return a set string?

    --Adam

  • @Kadar said:
    Maybe something to make sure port 25 is working for SMTP. Something simple to replace a lot of functions nagios checks

    I already have SMTP check, thanks! But I could add checking for custom answers, if that's what you mean?

    --Adam

  • @matthewvz said:
    Game server monitoring would be amazing. I would love a central place to see all of my player counts and uptime.

    What type of monitoring would this involve? Just port monitoring or a script running on the server checking things? Not really looked into game server side of things.

    --Adam

  • @jimpop said:
    I like this idea (please make sure you support ssh keys). Also, the ability to take action and execute commands via an ssh session. For instance, if latency is high to a server, I would like the remote ability to ssh in and shut it down or restart, etc.

    This is interesting. I had considered 'actions' on detected downtime. Connecting via SSH (or via IPMI, etc) is an interesting one. Do people really want their servers rebooted on detected downtime? Or I guess the system could SSH in and restart a specific service. Awesome suggestion, thanks!

    --Adam

  • Make it All in One monitor

  • @Kyn_DH said:
    Make it All in One monitor

    That's the idea. :-)

  • @adly said:
    Do people really want their servers rebooted on detected downtime?

    Consider a website with 3 servers (3 A/AAAA records in DNS). If latency to one host is extremely high, client browsers will still possibly reach that server and the user will experience poor performance. Using a monitor to detect extreme latency, and then take an action to shutdown nginx/apache, would cause the client browsers to only connect to the servers on the 2 remaining A/AAAA records.

  • @jimpop said:

    How about an option to connecto the Rage4/Amazon Route 53 to remove the entry from DNS? But, yes, I get what you're saying. Consider it added and once launched I'll send you a lifetime coupon unless you wish otherwise. :-)

    --Adam

  • @adly said:
    Consider it added and once launched I'll send you a lifetime coupon unless you wish otherwise. :-)

    How about a lifetime discount, I don't want you to create something awesome and then go out of business because of too many expenses. ;-)

  • @jimpop said:

    I'll link you a donation link, you pay what you think. ;-) I ain't gonna gonna retract my offer of a lifetime account, not into BSing people.

    Launch/beta Dec 1st 2013.

    --Adam

    Thanked by 1jimpop
  • Email notifications?

  • @Casius said:
    Email notifications?

    Yeah, got that. Plus Push/Mobile, SMS and HTTP(s). Can also do 'actions' as others have noted and adding now. :-)

    -Adam

  • mcmyhostmcmyhost Member
    edited November 2013

    @adly said:

    Maybe attempt SSH connection to check uptime?

    Maybe it would routinely run the

    top, free -m, tcpdump

    and report the results.

    Edit: Has been posted above.

  • @mcmyhost said:
    Edit: Has been posted above.

    That's not much different from pinging the SSH port. I was thinking connecting to SSH and actually looking for a response string and/or executing a command? Would that do?

    P.S. any and all suggestions are welcome, everyone brings a new perspective.

    --Adam

  • @adly said:
    --Adam

    An idea to prevent false downtime notifications would be.

    if - Ping Returns no response

    than - ping google.com to make sure server is online.

  • @mcmyhost said:
    than - ping google.com to make sure server is online.

    Yeah, that's a good idea. However, if one server thinks a server/host is offline it puts a request in to an urgent queue at which point 5 other nodes around the world do the exact same check. Additionally, before each batch of checks is run a node checks it's connectivity via a number of checks.

    The main aim of the project is to cut out false reports and notify only when 100% sure, but thanks for the suggestion!

    --Adam

  • Thought of another one..... Grouping: Allow me to "group" or tag a set of servers such that I will only be notified if x number/percentage of tests fail. For example: Suppose I have a web site with 3 A/AAAA DNS records (for 3 servers). I would like to monitor the 3 web servers, but I only care to be notified if 2 of the 3 servers fail a HTTP test (Note, I would also want to be able to be notified when any of the 3 servers fails a ICMP test).

  • Makkesk8Makkesk8 Member
    edited November 2013

    Free: 1 minute check's
    Paid: 10-30 sec check's

    The rest is pretty much obvious.

  • Is the focus of this a web interface? I'd like to be able to get a summary output easily from commandline, extended details + access to monitor settings would be even better.

  • An API would be a nice thing to have

Sign In or Register to comment.