Ask HN: Review my pivot, a server uptime monitor | Hacker News Reader

Ask HN: Review my pivot, a server uptime monitor

5 pointsJFitzDela7 comments

Okie dokie -- Monitor For Me as a home alarm monitoring service was an utter flop, so I let it stagnate while I moved on to other things.

The other day a critical server of mine went down and I only discovered it by chance. "Shazam," I thought, "a pivot is in order!" I reoriented the service, got it live, and then I discovered Pingdom, UptimeRobot, and the several other options (that I somehow missed in my single cursory Google search) that already provide monitoring and notifications.

Not to worry, though, because I still believe I can carve out a piece of the market.

Monitor For Me checks up to 10 websites (or any hostname/IP and port that can be opened) every minute per user for $15/month. It'll also capture and show you response times from (currently) three cities around the US, and I plan to add a 4th in the US and one in London before too long.

Obviously there are a tremendous number of reporting, connection types, and options in general that can be added for value, but I want to start with the MVP -- a simple service to let you know (by SMS or email) when your server goes down.

Any hackers around here mind signing up and trying it out and letting me know what you think? Am I wasting my time (because the product needs more oomph or because there just isn't market share available)?

Edit: Also, if you've used other services like this, what was missing? What was unnecessary?

If you don't mind taking a look, visit http://monitorfor.me (clicky below), and use the access code HNFTW (limited to 15 users, I'll update if it fills up) for a free month of access.

Thanks a million!

- John

This problem has been addressed over the last 20 years by more products and services than I can remember, ranging from csh scripts that would ping a site and send an email if it was down to HP OpenView.

There are more than a few and sub $100 tools for Windows that'll do this and a lot more, so charging $15 for a service isn't going to fly.

Further, to properly monitor uptime requires a lot more than just monitoring the host. You need to monitor the various services, and the health of those services. Is the web server returning a valid page or a 404?

Then there's the need to monitor hosts and services that are not publicly routable, getting into authentication or the need to run the service inside the firewall rather than via a public site.

I don't know if I agree with the notion that the service model itself (or at least it's pricing model) isn't viable....

People ARE using services like these (see UptimeRobot and Pingdom as the examples mentioned above). Some of the services are free, some cost $10 a month, some $40.

That said, MY implementation may not be viable, but it's not costing me much of anything to find out, so I'm still going to give it a go.

As to what the service offers, there is obviously more that needs to be done for critical infrastructure monitoring, but for the average Joe running a website or three, keeping an eye on the web server is a good, basic start.

I'm also thinking of people like affiliate marketers (which I'm also attempting to get into) that use their servers to host landing pages and ads (and their tracking) for which they may get charged without a result if the server is down. THESE guys just want to know immediately, I should think, that they may be losing a lot of money with no chance of ROI.

In any case, I certainly appreciate your feedback!

- John

So this works by pinging the server? I ask because a lot of downtime is the malfunction of a specific program or component that still keeps ping and http intact. How do you detect this?

And if it's a ping thing, why would I pay 15 bucks per month and trust a third party when I can write my own script to dump failed ping results to my email?

I fail to see the distinguishing feature that I need monitorfor.me for.

Thanks for the feedback. Obviously most of us in the highly technical crowd could very easily throw a script like this together, but I'm targeting a somewhat less technical crowd (still guys that know they need it, but don't program).

The other benefit is that I offer (rather, will offer -- don't have it replicated yet) geographically distributed, redundant checking for up to 10 sites. This is a bit harder (and more expensive) to accomplish for the general developer.

Regarding the method of detection, I actually use a standard "fopen," to the hostname and port you specify, rather than a straight ping. In this way you could test specifically for your webserver (port 80) or SMTP server (25) rather than a generic ping. I even use it to test custom socket server platforms I have running.

Does that answer your questions? Change your mind about the value of the service?

Thanks!

I'm sorry if I come across as crude, but in startups there's no room for sugar-coating.

Personally I think that anyone who feels they need this kind of service (e.g. a less-techy manager at a company) almost certainly has some go-to tech guy that could whip it up over lunch. If the company doesn't have such a guy, they're probably doomed anyway.

I just don't see much of a market here and I would be surprised if this thing, in its current iteration, would ever buy you ramen. You might be onto something, but I think you should aim higher.

Clickable: http://monitorfor.me

Invite code: HNFTW (on the registration page)

Not to pee on your cornflakes, but for basic ping monitoring, even geographically distributed, $15 a month is a VERY high price.

For that cost, I'd expect it to do a lot more.

Reply on news.ycombinator.com