Hacker News new | past | comments | ask | show | jobs | submit login
Ask HN: Please review my pet project - Ping Brigade
33 points by IgorPartola on Aug 4, 2010 | hide | past | favorite | 36 comments
Ping Brigade (http://www.pingbrigade.com/) is a webapp that lets you measure how fast your web server (and/or your network connection) is from several locations in the world at once. It started out as a side project but after spending several months working on it, it is now live. For the future, I plan on adding more locations around the world, a mobile version and a blog (to talk about various optimization techniques, review hosting providers, etc.)

Any feedback is greatly appreciated. Thank you very much.




Does the page load test include asset loading and javascript execution? Based on the numbers I just got, I'm guessing not. Google's Page Speed tool measures for these. If your tool did this in the simple interface you have and from around the world, I'd be hooked.

Definitely a great execution though. Clean design, good name, and cute logo.

EDIT: The reason it'd be great if the page load test included these is because a) I could then test for what Google is testing for when figuring my site's speed into search results [1], and b) I could easily test the real-world impact of using a CDN.

[1] http://searchengineland.com/google-now-counts-site-speed-as-...


It does not currently include asset loading, just loading of that URL. Tools like Page Speed and YSlow do this very well, but only from your current location. Combining Ping Brigade with those tools should give you the complete picture.

EDIT: But thank you for the suggestions. Certainly something to think about/add in phase 2.


The market of connectivity/website monitoring is very saturated at the moment, and it has been in this state for at least a couple of years. I spent several weeks on the market research for a very similar service and here is a snapshot of my bookmarks from that time:

http://i37.tinypic.com/2n6651g.png

This is just a subset of existing services, all of these are functional and useful. They target different users and have different pricing levels, but quite a few are actually free and there was two free services with over than 50 sensor points each. Some are simple, some are very (overly) comprehensive. Some are very aggressive marketed (Pingdom), some are low profile, but still sit at the top of respective Google searches.

So, in short :) - this makes sense as a hobby project, but trying to carve out a non-trivial market segment for it is going to be the challenge. The technical and design merit alone is not enough, it will come down to marketing, sales, support and knowing how exactly you want to position the service ("the bigger view"). Make sure you can justify the effort before committing to it. I know I couldn't.


Thank you for sharing this. I will definitely need to do more market research. As you said, right now this is just a hobby project. Hopefully, it will catch on in some niche. If not, at least I learned a lot from creating it.


Pretty clean interface. I understood what the site was about in a fraction of second and I tested a couple of website without any misunderstanding or anything in the design. Congrats on that.

+1 Because my girlfriend like those birds on the top of the page.

This might really be useful to system admins.

My suggestion:

- API (with good documentations!) to test from a specific country

- Use Notifo


Thank you. That was the goal: make it as clean and easy to use as possible. It seems that others that provide this kind of service have pretty clunky UI's. Glad your girlfriend likes the birds. That was a big consideration (they had to be approved by my wife).

I guess the API is the next thing I should tinker with.


Looks good, though open World tab by default. Didn't notice there were other tabs as well.


Very nice interface.

On the bottom of the results page, it says "Need a better server? Get one at Ping Control." but the "Ping Control" link appears to be broken.


Should be fixed now. Sorry about that.


* maybe when one has filled in the URL, tested the ping, if he checks the "page load" or "web server latency" button, you could load the results right away without letting us do the extra step to click "enter" again. * it would be interesting if you would keep the history per link and graph the performance at the different point in time where people tested the ping.


Yes, I struggled with how to have the user choose which test to run. In the end I went with something that would allow the page to work even if JavaScript is disabled. But I will certainly be revisiting this portion of it to minimize the number of clicks it takes.


I just ran it with NoScript FF add-on enabled, and I got the following:

Well, this is embarassing...

We have a bug here somewhere. If you think this is a big deal, please let us know.

Turning off NoScript, I got results.


Also, yes, I do keep a history of all the measurements. That is going to be a part of it.


Very nice! I like the simplicity of it as I've used other services as well like Global Net Watch.

A big deal for us would be having some Asia Pacific locations as that is typically the area with the worst performance for our site and the one we most like to keep a close eye on.

I'd also 2nd the ability to have a nice restful interface that returns some json back to us for some automated testing.

Very nice clean design!


I will be adding more locations ASAP. The trouble is often times finding reliable providers when I don't speak the language. That's why I've been focusing on US + Europe for now, but Asia and South America are next.


Nice - I suppose one feature would be to implement a RESTFUL api that returns JSON or XML results so it could be used programatically - sys admins could write cron jobs to test their servers etc...

or it could be integrated into a software build process / test framework (continuum, cruise control etc) and sites could be tested automatically upon deployment...


That is certainly another idea I considered. On top of that I could use this infrastructure to create a web monitoring service that is subscription-based.


Absolutely - people could create a 'named' suite of test/pings via a web interface, then trigger that named test or suite of tests via the API: pingbrigade.com/api/run/?userKey=XXX&suite=foo

...I would look for a business case before investing too much time! :-)


Hope this helps.

The service we use (globalnetwatch.com) will log in to our app using credentials we supplied, executes transactions we specify such as loading a specific page or bringing back results from a web service, tests the results(there is numeric data at this spot on the page, etc.) and emails us a daily status. It runs every 15 mins and alerts us immediately if there is an issue. We get to choose how many different locations they test from both within and outside the US.

Pings and page load are just the tip of the iceberg.

That's your competitor.

Best of luck


Interesting - Actually, I built Femtoo.com which can be used for a similar purpose! But I have JUST posted this article about my content extraction web service cQuery.com:

http://news.ycombinator.com/item?id=1574214

small world!


Thank you for pointing it out. I think there is a ton I could do with the infrastructure that supports this (multiple servers capable of doing HTTP and ICMP requests).


Love it! I'd recommend though that instead of using a hash in your URL, you use GET Data. So a URL would look like: http://pingbrigade.com/?query=google.com&type=1 . Could be useful in some situations.


I had it that way originally, but changed it for a couple of reasons: (1) That way you could create circular redirects very easily. My current setup uses POST so you can't directly create a loop. (2) That way these URL's are permanent. You can send them to your hosting provider when complaining about network/server issues.


It would be nice to have a score associated with the results, similar to Y-Slow. I don't expect at the early stages that you would provide guidance on how to improve, but showing my score vs. average, or best or something would be beneficial, so I know how badly I need to improve.


I got all requests failed on testing page load for http://beta.crowdeo.com

Also, I'd like to be able to test server speed by more than ping latency, i.e. actual throughput between various points.


Have you seen http://just-ping.com/ - not as nice an interface as yours but seems to have rather more locations covered.


No, I haven't seen it before, that's why I created PB. Thank you for the link. I'll keep an eye on what features they have that I don't.


Well, this is embarassing...

The page you were looking for was not found. Please make sure you have the correct URL. If you do that means we screwed up, so let us know.


Sorry about that. You must have caught me in the middle of a deploy. Should be all set now.


Small remark: your map of Europe appears to be strechted out along the Y-axis. Is that possible? Or is it just using an unfamiliar projection?


I am using whatever Google Charts API produces and at their maximum size (440x220 px). It may be that they use a strange projection.


clickable: http://www.pingbrigade.com/


very cool idea. I really like the logo :)


like the name


Thank you. The original name I was looking for was geometric.com, but a domain squatter has it and I don't want to deal with that on principle.


Pingbrigade brings it home for me




Consider applying for YC's Spring batch! Applications are open till Feb 11.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: