Over the past few days, London has been under siege from multiple rioting mobs of hooligans and thugs. GoSquared (based in London) has been enabling a number of sites reporting on the situation to understand how events have been unfolding in real-time.

London Riots Map, based on Google Maps was created when the riots broke out to chart the London riots from a first person perspective by plotting tweets with the hashtag #londonriots that have an associated postcode. Cravify, a service for mapping events has also been used in a similar way.

We decided to put together a simple visualisation of how the events in London have unfolded – both in terms of web traffic that we monitored on GoSquared (we provide real-time web analytics after all – perfect for events like this), and from our first hand perspective being based in London.

Read the rest of this entry →

GoSquared has always enabled you to understand what your visitors are doing on your site. Given that we show you who’s on your site right now, it’s always been tempting to “reach out” and assist people from the GoSquared dashboard.

Often you may notice visitors on your site who appear to be hitting pain points or who might be thinking about signing up to your site or service. And until now there’s been very little you could do other than make a note to improve it “for next time”.

All that changes today.

Start chatting to your visitors today.

GoSquared customers on the Plus plan and above can now pair GoSquared with one of the most popular live chat applications on the market, Olark.

Read the rest of this entry →

We’ve recently been running into a recurring issue with AWS’ Elastic Load Balancing service. The problem manifests itself in vastly increased latency or eventual dropout of requests against the load balancer.

We verified the health of our instances in each availability zone, and concluded that it was not a problem at the instance level. Second, we confirmed that our DNS queries were being routed to the ELB correctly. They were. So the issue was at the ELB level.

As it turns out, there appears to have been/still is a bug where ELB will suddenly decide that there are no available instances in a particular availability zone, or at least it will have trouble connecting to instances in that particular zone. Even if the instances are perfectly fine, and pass all health checks.

The solution was to try to identify the suspect zone, and de-register that zone and all of its instances from the ELB. Once re-registering the zone & instances completed, the ELB was back up to full functionality again.

Hope this helps next time you run into any ELB issues!

P.S. We’re big fans of Amazon Web Services – if you haven’t already, check out our recent feature on the AWS Blog.