Show HN: Erlywarn, Clever Uptime Monitoring. Any thoughts?
11 comments
Looks nice!
However, a couple of points:
* The £1/site/month is too low. I won't trust you to be able to rent and manage a network of decent machines for that price point, so I wouldn't trust your results. You're also going to eat costs trying to support cheap one-site-owners. Try a price point like "£5 for 5 sites a month!" so that even if you have one site, you pay at least £5 a month.
* Show me your (hopefully) geographically dispersed network. Like a map showing where your machines are with fancy lines and all that. I want to know you know what you're doing and at the very least you should have a network of independent machines, geographically spread. Tell me you're using independent datacenters/cloud hosting for maximum availability.
* I don't know the current state of Go, but as a potential customer I'm questioning the logic of using an experimental language (as I see it) for an uptime monitoring system. Presumably you want to have the highest uptime as possible, and maybe Go interpreters/compilers have bugs that will affect you. I would not advertise the fact you're using Go.
* Link me to a dashboard where I can see the statusses of your service, and any maintenance messages you might have had. This might strike you as odd, but I would trust you more if you had a public page where you can show past incidents and how you resolved them, or at least a table full of green lights showing all your stuff is up. Some things can fail (without losing uptime), and I'd like to see you being open about it.
I understand you're just beginning so cherry pick from this comment what you like :) Either way it's a nice looking landing page.
However, a couple of points:
* The £1/site/month is too low. I won't trust you to be able to rent and manage a network of decent machines for that price point, so I wouldn't trust your results. You're also going to eat costs trying to support cheap one-site-owners. Try a price point like "£5 for 5 sites a month!" so that even if you have one site, you pay at least £5 a month.
* Show me your (hopefully) geographically dispersed network. Like a map showing where your machines are with fancy lines and all that. I want to know you know what you're doing and at the very least you should have a network of independent machines, geographically spread. Tell me you're using independent datacenters/cloud hosting for maximum availability.
* I don't know the current state of Go, but as a potential customer I'm questioning the logic of using an experimental language (as I see it) for an uptime monitoring system. Presumably you want to have the highest uptime as possible, and maybe Go interpreters/compilers have bugs that will affect you. I would not advertise the fact you're using Go.
* Link me to a dashboard where I can see the statusses of your service, and any maintenance messages you might have had. This might strike you as odd, but I would trust you more if you had a public page where you can show past incidents and how you resolved them, or at least a table full of green lights showing all your stuff is up. Some things can fail (without losing uptime), and I'd like to see you being open about it.
I understand you're just beginning so cherry pick from this comment what you like :) Either way it's a nice looking landing page.
RE price. I found the "plans" where you pay for things you didn't need rather off-putting from other services, especially in our pay-for-what-you-eat CPU/bandwidth/etc "cloud" (cringe) world. Higher price -> trust is a tricky one.
RE network & dashboard: Good plan. It does seem silly not to show we are using it too.
RE go; go is really not experimental anymore, and has been an excellent choice for this kind of concurrent problem. We had something similar in Erlang (internally) and this has been so much faster to build and is much simpler. I get your point and wouldn't normally mention it, but if anyone would be interested it's here :)
Thanks for your feedback.
RE network & dashboard: Good plan. It does seem silly not to show we are using it too.
RE go; go is really not experimental anymore, and has been an excellent choice for this kind of concurrent problem. We had something similar in Erlang (internally) and this has been so much faster to build and is much simpler. I get your point and wouldn't normally mention it, but if anyone would be interested it's here :)
Thanks for your feedback.
You're using Go...but...it's ERLywarn. So how can it possibly be written in anything but Erlang. See ErlyWeb, Erlware, etc.
I'm sorry but the name of your product needs to change to Gowarn. I don't make the rules, I just enforce them.
(Best of luck, looks good and not overpriced like some freemium alternatives)
I'm sorry but the name of your product needs to change to Gowarn. I don't make the rules, I just enforce them.
(Best of luck, looks good and not overpriced like some freemium alternatives)
Last time I checked, the price separator that other monitoring sites used -- in particular between free and paid accounts -- is frequency of checks. E.g., free accounts include checks every half hour (or 15 min), and paid accounts could be as often as every 30 seconds.
With that in mind, the frequency of checks (or: "how long will my site be down before you tell me?") is an important thing to mention; I couldn't find that info, scanning through.
The pale grey text costs some effort to read, as well, so I'd also suggest increasing the contrast there.
Looks like a good site, and simple, though it may prove a bit too easy for the major existing monitoring sites to copy your distinguishing feature. Good luck!
With that in mind, the frequency of checks (or: "how long will my site be down before you tell me?") is an important thing to mention; I couldn't find that info, scanning through.
The pale grey text costs some effort to read, as well, so I'd also suggest increasing the contrast there.
Looks like a good site, and simple, though it may prove a bit too easy for the major existing monitoring sites to copy your distinguishing feature. Good luck!
Looks great, always good to see a UK based company innovating.
I've been using wasitup.com and just the other day I was thinking why aren't there any reasonably priced monitoring services for small single site owners? You've filled that gap and innovated at the same time. Congrats, I wish you every success.
Think you have a typo, search about section for Earlywarn instead of Erlywarn.
I've been using wasitup.com and just the other day I was thinking why aren't there any reasonably priced monitoring services for small single site owners? You've filled that gap and innovated at the same time. Congrats, I wish you every success.
Think you have a typo, search about section for Earlywarn instead of Erlywarn.
eh, I'd ignore people who say that low prices make you look less reliable. Being new makes you look less reliable. In the infrastructure market, it's traditional to start at the low end and work your way up, and SysAdmins are some of the cheapest people you could ever meet, so having a fully functional low-end product is a good thing. If you want a reputation, you need to earn it.
That said, SysAdmins are also an inquisitive bunch. As someone else pointed out, you need to document where the test servers are, and what happens when one test server can reach the site but another can not.
If you only have one location doing the checking, or if you can't handle the case of the site being accessible from one place but not another, well... that's a very important feature. Additional servers are cheap; but you will have to do something sensible when one monitoring site says it's up and another says it's down, and that's a moderately difficult interface problem you need to solve (and one that most monitoring software does badly.)
That said, SysAdmins are also an inquisitive bunch. As someone else pointed out, you need to document where the test servers are, and what happens when one test server can reach the site but another can not.
If you only have one location doing the checking, or if you can't handle the case of the site being accessible from one place but not another, well... that's a very important feature. Additional servers are cheap; but you will have to do something sensible when one monitoring site says it's up and another says it's down, and that's a moderately difficult interface problem you need to solve (and one that most monitoring software does badly.)
I tend to agree on this pricing comment -- being new means no one will have much faith you'll still be online a year from now. But you're cheap enough that if you look interesting enough, lots of folks will add you as a backup monitor.
From that point your foot is in the door, and it's up to you how to find the leverage to become their primary monitoring service. You can raise the price (possibly for new fancy features only) as you develop more features and get some history of reliability behind you (which is a feature in and of itself, of course...).
From that point your foot is in the door, and it's up to you how to find the leverage to become their primary monitoring service. You can raise the price (possibly for new fancy features only) as you develop more features and get some history of reliability behind you (which is a feature in and of itself, of course...).
Clickable: http://erlywarn.com
Typo on the front page: "Recieve"
Typo on the front page: "Recieve"
ah thanks... I seem to have lost spellcheck in vim
Reading that text on the website in light gray was kind of painful. I get that you need to differentiate some blocks of texts but those lines look like they are fading.
Oh, and maybe a lightbox with prev-next buttons? Nice work tho, best of luck.
The site looks good, although I prefer the low price you may be signalling low quality with 1£.
BTW, it's nice to see that twitter's bootstrap framework is being used by alot of new projects. They sure filled a gap for non-designers.
I'm actually not too bad at design, and really enjoy it – my other startup http://www.pacpacs.com was designed by me which I don't think is too bad :-)
... where bootstrap really shines is that I had a working basic version of the app running and looking pretty good about 6 hours after the idea, I would have spent longer than that in photoshop staring at color swatches. It's the perfect prototyper.
... where bootstrap really shines is that I had a working basic version of the app running and looking pretty good about 6 hours after the idea, I would have spent longer than that in photoshop staring at color swatches. It's the perfect prototyper.
Looks good, although takes ages to load. Having design and coding skills is a dead combo! Keep up the good work.
Purchased! We already use Pingdom and NewRelic so feel free to get in touch for feedback and use case support. First comment is you need to add https support to the adding a domain UI.
Thanks! We'll be making lots of improvements over the coming weeks so if you feel anything is missing drop me a note.
yeah I had though about https not being clear, it will accept your https url, but your right there's no indication that it's ok to type it.
yeah I had though about https not being clear, it will accept your https url, but your right there's no indication that it's ok to type it.
nice service with a frictionless signup. very cool.
what does this do that keynote does not?
keynote is aimed very much at monitoring and fine tuning a single large application or organisation. It's very powerful but also quite complex. So isn't really suitable for someone with little technical knowledge.
We're aiming for more of a fire-and-forget approach where Erlywarn learns what is acceptable/unacceptable behaviour – next week the notification emails will have a "don't warm me like this" type button in them to help you train the alerts. When monitoring 100's of sites you have some that are more sensitive than others, and we want to remove the "setup" from this.
We're aiming for more of a fire-and-forget approach where Erlywarn learns what is acceptable/unacceptable behaviour – next week the notification emails will have a "don't warm me like this" type button in them to help you train the alerts. When monitoring 100's of sites you have some that are more sensitive than others, and we want to remove the "setup" from this.
This looks really cool.
As a student I love your pricing of a buck per site, but as relix has mentioned, that might seem too cheap to a lot of people so a small package of a few websites would also work.
What happens if for whatever reason the site only starts behaving erratically for a class of users? Say, my servers are in Europe and the whole of US can't access my site, but it all looks fine from Europe?
As a student I love your pricing of a buck per site, but as relix has mentioned, that might seem too cheap to a lot of people so a small package of a few websites would also work.
What happens if for whatever reason the site only starts behaving erratically for a class of users? Say, my servers are in Europe and the whole of US can't access my site, but it all looks fine from Europe?
Thanks.
At the moment that would be flagged as "problem" and you would be alerted. Very soon (few days) the notifications will be much more informative and tell you exactly where and what kind of problem.
At the moment that would be flagged as "problem" and you would be alerted. Very soon (few days) the notifications will be much more informative and tell you exactly where and what kind of problem.
http://erlywarn.com
The project was scratching an itch that monitoring services would only alert you if your site went down and not if they started responding slowly or erratically.
So we built Erlywarn to keep track of the response times and allow it to send notifications if your site responds unusually, as this is often an early warning sign that something is starting to go wrong.
It's only been 2 weeks from conception to this, so feedback would be most welcome, especially:
Would you consider using the service yet, what are we missing?
Prices are currently in GBP (while we sort out a merchant account) is this a big turn off for the US?
Would $1/per site/month we good value to you?
Thanks for any help
ps: written in Go for those that find that sort of thing interesting.