A regular theme I talk about with vendors is the move for what could formerly be though of as purely point solutions into a more holistic view of the world. The big opportunities in the cloud revolve around finding functions that can straddle a variety of workloads, applications or other silos and more and more vendors are hearing this message and considering how it applies to them. A case in point is blitz.io, an application testing service that I’ve written abut before. I caught up with Kowsik Guruswamy, CTO of Mu Dynamics, the parent company behind blitz.io to see what this holistic view of the world means for them.
First an operating update from the service. Blitz.io tolkd me of a variety of metrics that speak to their success;
- Half a billion hits since launch
- 80000 performance tests run so far
- 14000 users (a doubling over the last three months)
- Well integrated in the PaaS ecosystem including integrations with Engine Yard, CloudControl and AppFog
All of which speaks to blitz.io being more than simply performance testing. It is interesting to note that Guruswamy reports that there are very few performance engineering people using the application – rather their main user base are developers, IT departments and systems administration folks – a range of users across an application’s lifecycle. So how has blitz.io broadened its base, and what does this mean for the broader ecosystem?
Blitz.io is now integrated with NewRelic, the monitoring solution, this integration gives users not only a forward looking view of performance (from the load testing perspective) but also a real time insight into application and infrastructure bottlenecks. This takes what could be regarded as a solution to be used primarily at one point in an applications lifecycle, to one that is deeply embedded in the day to day running of a workload.
Blitz.io have also moved to an hourly pricing model for its paid offering – this move is encouraging users to look at load testing as a tool to be used regularly and often, and it is also democratizes load testing for users who might want to use it in short bursts.
Finally blitz.io is introducing its own basic monitoring service – while this could be considered competitive with the integrations of other monitoring services, it speaks to my contention that anyone building an application, infrastructure or networking management tool needs to think holistically, and embrace the breadth of functionality that users might need. In terms of what they’re offering, blitz.io can already do transactions (cookies, HTTP verbs, multi-url’s, parameters, variables, etc). They are now allowing customers to run these concurrent transactions periodically. This simulates “normal” load on a site and they then use the response time to measure the apdex score. If this score falls below a certain threshold, the customer can be alerted through email, pager-duty, campfire, etc. The service can also be run across eight different regions globally to simulate local conditions.
In the move from traditional models to a DevOps/NoOps paradigm – tools like blitz.io will become more important – the company is positioning themselves well for the future.