Nice job for releasing this. However, seems hard to compete with pingdom, who have exponentially more features and integrations and are relatively inexpensive. Kuddos on shipping!
Thanks! No intention to compete, as it says on the page it's not meant for professional purposes (of course, success might change that). It's mostly a pet project to monitor pet projects.
As a DevOps, I'd like to disagree with this. Pingdom previously was a joy to work with. Now that they're attempting to bring in PagerDuty-esq features (escalation policies instead of simple alerting), their UX leaves a bit to be desired. Also, I don't have high hopes after the Solarwinds acquisition.
It's not as simple as you'd think. Monitoring 1 website is easy, monitoring 1M websites and gathering results in realtime from stations around the globe to make a proper decision is a different thing. Significant bandwidth, and CPU are needed to achieve this, not to mention relying on third party services like twilio and sendgrid to notify customers and that cost money too.
I've been looking for a pingdom replacement for the very same reasons GP mentioned. Webmon does look interesting, even though it does appear a bit more expensive. It seems penny-pinching if you consider the cost of downtime, but despite all my complaints about pingdom, it still works and does its job. So I'd need a compelling reason to switch and it's always easy to convince upstream when the price is cheaper or you get something better.
Just a quick question: what's a 'target'? is it the equivalent of a pingdom 'check'? or is this an endpoint IP that we can run several checks over?
Some feedback after a quick glance on the homepage - it's nice and seems to provide the most important info, but I'd appreciate some more in-depth info on some areas (to answer my question about targets and possibly other questions about other stuff: what's a user, what's a dashboard etc)
btw, the ping engine is a Go app that pings each hosts in a separate goroutine. I might open source it in the future once the code has more tests and less shame.
I wouldn't worry too much about it. Everybody writes horrible duct-tape code at some point, and showing how you improved/refactored the code is a much nicer thing than showing how you supposedly magically got it right the first time :)
thanks. Indeed the service totally lacks features and maybe reliability, but my whole intention was to find the lowest possible barrier of entry, which I too find very high in most online tools.
This is awesome but what is the motivation behind this? Why do you care to set up a server and pay for the (minimal) bandwidth that is needed for this service? Is this purely an act of goodwill? Do you hope that this will gain traction so that you can start charging for the service? Are you secretly harvesting clusters of IP addresses of HN readers?
Good question! I think most of us with pet projects do it either to learn something or with the hopes of gaining a lot of traction and one day implement some paid feature. For me it was both.
Maybe one day if there are a lot of users I'll implement paid accounts with more than 10 alerts and history dashboard. But most likely it will never gain enough traction for that, in which case it's no problem, since I have 6 tiny servers for different activities and this only uses 2 of the with minimal resources (it's just ping). It will just sit there alerting me and some others when something goes down.
Thanks! "Trust" on the internet is never a thing and I have been growing suspicious of it more everyday. Way back when you could almost trust your DCC on IRC but lucrativeness involved with the 'unwashed masses' in using the internet make this horrible problem for a technologist exploring new things.
+1 to this. I've set up a few services of a similar nature, and a couple were worth enough to me to go to a further length and monetize, but only insofar as it paid my server bills and added a little on top for taxes and such.
It feels great to hack on something like this, and not worry about billing integration and supporting paid customers at first.
hahaha, how ironic. No, there is no meta monitoring. The site it's actually up, just really slow ATM, but I'm logged it and everything look fine and access logs as fluent. I think it might be some connectivity issues with digital ocean.
actually, I just realized someone is trying to DOS the site creating massive random accounts. Please unknown hacker, don't be a dick, this is just a pet project.
Any plans to implement feature like complete server monitoring? Like Google analytic's for Server monitoring?
This is simple yet helpful tool. Those who wish to stay away from nitty gritty of configuring complex apps to monitor their server this can be helpful.
Indeed there are plans in my head. But that could only happen if I manage to get a big user base while at the same time I find a way to fight abuse without compromising simplicity.
Congrats on shipping. I run a very different kind of monitoring service (alerts when your cron jobs and other scheduled tasks don't run or take too long, see my profile for details). I'm wondering why you don't offer paid plans. In addition to reward for your work, business users really would rather rely on a business and not a hobby. Maybe your target market is hobbyists but you should certainly consider a paid plan early on. Our subscribers are an important voice in our product decisions and doubly so early on.
Thanks. This project for me is just a proof of concept. Having paid customers would require dedicating massively more time to initial development and maintenance, which simply I'm not ready to do until the idea is somewhat validated. Best of luck with your venture!
Maybe you should slow down the ping interval a bit, because there is no authentication and someone might run it on a big number of IPs of the same network class, and that will quickly get your IP banned on routers. No one really needs to check the server every 5 seconds (not on free service at least). Perhaps you can make a burst of 4-5 pings every minute or 5 minutes, that would be quite sufficient IMHO, and would significantly reduce your traffic and expenses.
I gave the interval some thought, and while I might have to increase it in the future, I think 5 seconds is a good place to start. This is just a single innocuous ICMP echo packet, not port scanning. And a single user can only set up 10 pings, so it will require quite some manual effort to generate some noticeable traffic.
When I implement HTTP checks however, yes, the interval will have to be in the minutes.
I've never been block by just pinging anything, but if that happens for someone, well... it'll simply won't work for that user. No big deal, that's the upside of being a free no-SLA service :)
As for my own side of the net, I asked someone at DO and I was told that it should not be a problem.
if someone knowns your email address and that you've confirmed it, then yes, that someone could set up more, but the limit is 10.
If that ever happens to you, you could:
1) pause the unwanted alerts, and remove some only when you need new ones.
2) if you panic, use the "unsubscribe" button that will block any future emails to your address.
Having such a low barrier of entry does not come without tradeoffs.
$ curl ping.gg/monitoring@me.com/127.0.0.1
[ ok ] Ping alert created for host '127.0.0.1' and email 'monitoring@me.com'. Check your email to activate it.
64 comments
[ 2.9 ms ] story [ 135 ms ] threadAs a DevOps, I'd like to disagree with this. Pingdom previously was a joy to work with. Now that they're attempting to bring in PagerDuty-esq features (escalation policies instead of simple alerting), their UX leaves a bit to be desired. Also, I don't have high hopes after the Solarwinds acquisition.
Just a quick question: what's a 'target'? is it the equivalent of a pingdom 'check'? or is this an endpoint IP that we can run several checks over?
Some feedback after a quick glance on the homepage - it's nice and seems to provide the most important info, but I'd appreciate some more in-depth info on some areas (to answer my question about targets and possibly other questions about other stuff: what's a user, what's a dashboard etc)
BTW, +1 to HTTP support, a must-have.
Emails look like this: http://i.imgur.com/6kFqSnm.png
+1 noted!
and that's the _good_ part.
Maybe one day if there are a lot of users I'll implement paid accounts with more than 10 alerts and history dashboard. But most likely it will never gain enough traction for that, in which case it's no problem, since I have 6 tiny servers for different activities and this only uses 2 of the with minimal resources (it's just ping). It will just sit there alerting me and some others when something goes down.
It feels great to hack on something like this, and not worry about billing integration and supporting paid customers at first.
This is simple yet helpful tool. Those who wish to stay away from nitty gritty of configuring complex apps to monitor their server this can be helpful.
I wrote a software like this, tried to ship it but I failed. Now it's open-source for everyone.
If you're interested, grab it.
https://github.com/rafaqueque/responsly
When I implement HTTP checks however, yes, the interval will have to be in the minutes.
Every 5 minutes is more than fine for a free service (and do future upgrades for every 1 min or constant monitoring).
In any case let the user select their interval.
Nice work btw!
As for my own side of the net, I asked someone at DO and I was told that it should not be a problem.
If that ever happens to you, you could: 1) pause the unwanted alerts, and remove some only when you need new ones. 2) if you panic, use the "unsubscribe" button that will block any future emails to your address.
Having such a low barrier of entry does not come without tradeoffs.
$ curl ping.gg/monitoring@me.com/127.0.0.1 [ ok ] Ping alert created for host '127.0.0.1' and email 'monitoring@me.com'. Check your email to activate it.
There is an alert limit per host, so not everyone will be able to do the same. Congrats!