Freelance Web Consultant at architxt.net
23 January 2006 12:46pm
Does anyone know of UK based online service/tool that regularly checks for website availability?
Something along the lines of: http://www.alertsite.com/web_site_monitoring_suite.shtml
CEO at Econsultancy
23 January 2006 13:33pm
We use Site Confidence - you didn't specicfy whether you were after a free service or not. They're not free but we've found them very good and the service does a lot more than just basic uptime monitoring.
23 January 2006 13:43pm
Thanks - looks like a good option.
It also occured to me that there may be a desktop app that does the same. Nothing too sophisticated - just regular checks that the site is live and email alerts when it isn't.
Managing Director at Box UK
24 January 2006 21:20pm
We have used site monitoring tools for some time, and a quick search on Google shows that there are now a plethora of products available. I haven't come across a desktop app which does this although they no doubt exist. One issue is that its usefulness would depend on the uptime of the desktop and local ISP etc.
Site monitoring is also one activity which you can happily look abroad for (and thus save via the exchange rate/market competitiveness). This may also mean cheaper SMS alerts (this is certainly our preferred option although be prepared to hear your phone beep in the early hours...).
CEO at SciVisum.co.uk
30 January 2006 10:53am
Make sure you've thought about what you want to monitor.
Homepage monitoring on it's own for example is of little value.
Your homepage may be up, but your online store or login systems or etc could be dead...!because they depend on different technology code and databases than the home page.
We do a load of web monitoring and stress testing in the UK here at SciVisum, and it's true 80% of the time that a web site application will have some areas that perform much much worse than the rest.
So instead of thinking page-monitoring, you need to be thinking User-Journey-monitoring.
Work out what the important features on your site are - they're the things that are making the money, or servicing your clients. User Journeys are the multi-page routes that visitors take to make use of those features.
Monitoring those user Journeys will then be the vital KPIs that you can pass to the powers that be, as evidence of what user experience visitors are getting.
When the call centre says that people complained about the web being broken yesterday, your User Journey graphs give the hard evidence as to whether it was the website, or your customer's own PCs or ISPs.
When your media buying brings in lots more visitors but at a lower sales conversion, the Journey graphs will provide the evidence to know whether performance issues on the site may explain the lower conversion rate - so don't blame the marketing campaign!
Not only does monitoring a range of User Journeys provide vital KPIs, it also provides a common language between the business folk and the tech teams. Hard evidence graphs of each journey will show whether there's something the tech team need to fix, or whether they have been blamed unfairly.
Getting the right choice of user journeys is key - and getting both the business and tech teams to buy into the selection of journeys is the first step to getting attention back on your customers experience.
Also when selecting monitoring suppliers, don't forget that monitoring is more than just uptime/downtime.
Monitoring should give you graphical data of performance timings - just how long does it take a user to go down the X page route from homepage to add-to-basket? How slow does it get Friday evenings or etc?
And monitoring is also going to highlight whether any of the Journeys on your site suffer from sporadic errors. These are brief outages, that are not matched with anything the tech team can relate to - no server crashes or CPU overloads or memory swapping. They happen infrequently but throughout the week, and more often during busy times. They don't go away when hardware is upgraded...
They are most often due to design problems in your application code (race conditions), or the database locking area. If your call centre keeps reporting customers saying your web site is broken, and your tech team report no 'outages'... then you've likely got sporadic errors.
I'd say that 1 in 5 of the user journeys that we look at suffer (more than 1% error rate).
Be aware when choosing a monitoring supplier - that running at a fixed frequency of say 5 or 10 minutes is pretty ineffective at finding sporadic errors.
Look for a service that increases the sampling rate when an error is detected - eg 5 minute sampling increases to 30 seconds. This way, the classic sporadic error profile (errors are nearly always just one sample) is easily visible, and can be seperated from brief server outages.
Another thing to consider, now that about 60% of UK users are broadband, don't go with a supplier that graphs performance for modem-speed users. The graphs will pretty much just be an exact correlation to the size of your pages... for a modem user the 5Kbyte/sec bandwidth limit is the bottleneck not your servers. So unless your servers are *really* struggling and can't even keep up with that, then you'll just see a flat line that only changes when your page sizes change. Not much value at all.
Lasly, you mention cheaper SMS alerts?
Well, if your site is generating revenue or supporting your customers, it's worth spending a commensurate amount on monitoring too, so surely a few 5p text messages is neither here nor there!
Anyway, you can choose a supplier like us that includes SMS messages in at no charge. (we're not at the $50/month service cheap & cheerful end of the market of course.)
Phew, didn't mean this reply to wander on so long...
...think I'm feeling cynical this week about the inability of so many organisations we look at to run a web site with sensible performance. Just yesterday a client was running through the user journeys they were interested in. Bang, journey couldn't get any further, the server just pushed us back to a previous page. No explanation text. "Don't worry" says the client, "it's a database thingy-run, happens 3 times a day and when it runs that element on the site doesn't work. We'll have to wait 5 or ten minutes..."
Hopefully our monitoring of this global player's various key journeys will provide the hard evidence, to get their suppliers (being paid enormous 6 figure sums) to fix this, and maybe a bunch of other issues that they're not aware of right now too.
DeriSciVisum - web monitoring
Free market research on digital marketing
Daily Pulse: award winning newsletter
It takes 30 seconds to register