Head Guru Posted November 29, 2004 Posted November 29, 2004 Hey gang! Hope everyone had a nice weekend. I am working on a new server monitoring page and would like to ask a couple questions. What would users prefer. A page that give a lot of information. Response times, packet loss, round trip information and so on. Pretty detailed information. However, the page would take a bit of time to load and you would have to scroll down a ways to get to your server. OR A page that simply tells if the server and services are up or down. Any input would be helpful. Thanks
Madmanmcp Posted November 29, 2004 Posted November 29, 2004 Sometimes too much information will cause more headaches than its worth. I can see the concerned users now wanting to know how to figure out exactly what part of his data was "lost" in the 4% packet lost field Personally I would perfer all the data you are willing to give me and I'll ignore what is not needed and use what is.
TCH-Rob Posted November 29, 2004 Posted November 29, 2004 From a users perspective all I want to know is if it is up or down. If it is up then I can try my own diagnostics to see where the failure is occuring. I also think Bob has a point in that too much information may confuse those that dont know what it means to begin with.
TCH-Don Posted November 29, 2004 Posted November 29, 2004 For me a page that shows the server is working is fine. if I want more info I can goto my cpanel and check on the server.
Virtual Imager Posted November 29, 2004 Posted November 29, 2004 Why can't you do both? Have a page that simply tells if the server is up or down, with a link to another page with more information. Just a thought...
AlanV Posted November 29, 2004 Posted November 29, 2004 Both would be great, but if I had to choose, I'd prefer something that simply said if things were up or down.
TCH-Thomas Posted November 29, 2004 Posted November 29, 2004 Agrees on Bob, Rob and all the others. If its too much that is presented in one page, its confusing, so I agree on VI on that. But instead of having one page saying the server is up or down and then a link to the rest of the info, make it several links to several pages handling one thing/page with explanation what each presented thing means. This way the user doesnt only gets educated but will find out if blahblah is very much trouble or just a tiny trouble. Ok, long explanation, but I think you know what I mean.
arvind Posted November 29, 2004 Posted November 29, 2004 I'd prefer both too, but if its an either/or situation, I'd prefer simply up/down
Head Guru Posted November 29, 2004 Author Posted November 29, 2004 Keep the input coming. I think it will be either or in this case. We already offer tons of server status information on our status site. http://www.tchstatus.com I am upgrading our main server status page. I think the current one works, but is lacking on real information. For example. If I were a user, I would like to see how many failed tests occured over the past 24 hours. I will be posting some beta versions of the new status page shortly. Thanks for the input and keep it coming.
HCSuperStores Posted November 29, 2004 Posted November 29, 2004 If it were possible, I would go with the both option. A page that quickly tells you if there is a problem where you can find your server quickly and easily. If the status page shows my server's HTTP services are down, I don't bother to look any further. However, if the server is working, the additional detail would give troubleshooting information for the technically minded people out here. Additionally, it would add to the atmosphere of "transparency" that we have continued to love and enjoy as being part of TCH. Being able to talk with support about and point to a specific problem with the server may help get things working more quickly. Just my thoughts ...
TCH-Bruce Posted November 29, 2004 Posted November 29, 2004 If I were a user, I would like to see how many failed tests occured over the past 24 hours. While you or I would want to see information like that most of the users would be baffled by it. Simple is it up or is it down is all that is necessary for the most part. Being able to link to more detail would be great.
curtis Posted November 29, 2004 Posted November 29, 2004 Simple is it up or is it down is all that is necessary for the most part. This gets my vote I could care less why, just if it is or not.
dcumpian Posted November 29, 2004 Posted November 29, 2004 How about up or down and automatically submits a support ticket when down for any users on the server? Regards, Dan
btrfld Posted November 29, 2004 Posted November 29, 2004 I'd vote for all the information. It should be fairly apparent whether the server is up or down on the whole. As BoB says, one can decide which information to ignore.
TCH-Rob Posted November 29, 2004 Posted November 29, 2004 As BoB says, one can decide which information to ignore. Provided you know what that information is for.
TCH-Dick Posted November 29, 2004 Posted November 29, 2004 How about up or down and automatically submits a support ticket when down for any users on the server? Regards, Dan Between all the monitoring we do auto submitting a support ticket would be a waste of resources for the most part We have our own monitoring system that pages Bill, Alan our lead sysadmin and myself. We also have a 3rd party that does monitoring from outside the network that also pages all of the above. Alan also has his own monitoring that alerts him as well as the Sr. techs. Bill runs server monitoring from a server at his location and I also run a monitor system from my house as well. Bill and i both carry blackberrys pretty much 24/7 that get sent a detailed e-mail every time a service doesnt respond. Sooo, we usually know about an outage before anyone else does
Madmanmcp Posted November 29, 2004 Posted November 29, 2004 How about up or down and automatically submits a support ticket when down for any users on the server? Thats already part of monitoring. The "ticket" is probably manually opened by the Data Centers monitoring group. There is also probably an automatic email generated and sent to Bills pager when the problem is serious enough (server down) As far as sending something to "any users on the server" thats stretching it a bit. They do send out for down times that are for extended periods but down times are mostly a temp problem and it comes right back up. Notification for these would be unnecessary.
dcumpian Posted November 29, 2004 Posted November 29, 2004 How about up or down and automatically submits a support ticket when down for any users on the server? Regards, Dan Between all the monitoring we do auto submitting a support ticket would be a waste of resources for the most part We have our own monitoring system that pages Bill, Alan our lead sysadmin and myself. We also have a 3rd party that does monitoring from outside the network that also pages all of the above. Alan also has his own monitoring that alerts him as well as the Sr. techs. Bill runs server monitoring from a server at his location and I also run a monitor system from my house as well. Bill and i both carry blackberrys pretty much 24/7 that get sent a detailed e-mail every time a service doesnt respond. Sooo, we usually know about an outage before anyone else does How do you ever get any sleep? I handle on-call support for the company I work for and it can be a royal pain... Regards Dan
dcumpian Posted November 29, 2004 Posted November 29, 2004 How about up or down and automatically submits a support ticket when down for any users on the server? Thats already part of monitoring. The "ticket" is probably manually opened by the Data Centers monitoring group. There is also probably an automatic email generated and sent to Bills pager when the problem is serious enough (server down) As far as sending something to "any users on the server" thats stretching it a bit. They do send out for down times that are for extended periods but down times are mostly a temp problem and it comes right back up. Notification for these would be unnecessary. You are right that if the problem is brief then I, as a TCH customer probably wouldn't care, but it would be nice to know when there is a problem lasting longer than say, 15 minutes. Regards, Dan
KungFòóFairy Posted November 29, 2004 Posted November 29, 2004 My choice would be option number one: more detailed information. :Nerd:
Madmanmcp Posted November 29, 2004 Posted November 29, 2004 would be nice to know when there is a problem lasting longer than say, 15 minutes A 15 minute SLA is standard in the industry ... and they pay a very high price to get it! Here in the shared server market where you are charged the price of a Happy Meal for service...I don't think so . I'll be happy just to see a server status available in real time on the web.
ace Posted November 30, 2004 Posted November 30, 2004 In my opinion, for the main status page I would keep it simple. I would like to know when services are down on the server or problematic. May I recommend three colors. Green for all systems "OK" Amber for services that are having "issues" and Red for services "offline" Sometimes a service may still be running but running crippled. May I suggest using BigBrother for a monitoring systems details. That is what we run in our data centers. I think details have their place but not on a main status page. Just my 2 cents.
ace Posted November 30, 2004 Posted November 30, 2004 Bill and i both carry blackberrys pretty much 24/7 that get sent a detailed e-mail every time a service doesnt respond. Sooo, we usually know about an outage before anyone else does Did someone say Blackberry's? We....Haaaaa way to go Bill and Mike. What a great choice for a moblie device!! Sounds like I work for the company that makes the Blackberry...........Oh! that's right, I guess I do. All joking aside I think that the Blackberry is a great mobile communication device. I hope they work well for your desired needs.
taznumber1 Posted November 30, 2004 Posted November 30, 2004 I say give both where the user can click on lets say the server name to get the more detailed information about there server, but pretty much it would be nice for a quick glance to see a green, yellow, or red button to say if the server is good to go or not.
dcumpian Posted November 30, 2004 Posted November 30, 2004 would be nice to know when there is a problem lasting longer than say, 15 minutes A 15 minute SLA is standard in the industry ... and they pay a very high price to get it! Here in the shared server market where you are charged the price of a Happy Meal for service...I don't think so . I'll be happy just to see a server status available in real time on the web. Probably true. But "shoot for the moon" and call it a success if you get into orbit Regards, Dan
Head Guru Posted November 30, 2004 Author Posted November 30, 2004 I think we will offer a status page as it is now, but there will be a link near the top of the status page "Click here for a more detailed report". This will give the best of both worlds? Bill
ace Posted December 1, 2004 Posted December 1, 2004 I noticed that the IMCP & myql on the status page show 0ms in reply time. Is this correct? (see attached) stats.doc
Recommended Posts