For a few months now our infrastructure engineering team members have been working on a major upgrade to Accelo. With code and server architecture changes across our entire stack, this upgrade is designed to provide a more scalable, responsive and, above all, fast user experience for Accelo.
The best way to think about this upgrade is like the change between a DC-10 and a Dreamliner - they're both wide-body jet airliners that fly, but the Dreamliner is faster, more efficient, carries more people in more comfort and is of course a lot more modern.
With system changes like this, the best result is no-one noticing anything change except a speed improvement - everything needs to keep working just the same as it was under the old architecture (just better). With literally thousands of lines of code changed across the entire product, however, there's still likely a few wrinkles to iron out - and we need your help.
The good news is, you can start giving this new version of Accelo a test-flight today if you like - simply change your deployment URL by adding "alpha" to your domain before the .accelo.com bit and log right on in.
When you're using the new Alpha platform you're still using all of the same Accelo data (database, file store, activities & timesheets, etc) as in your main production system. This is simply a new, faster front-end interface into your same live, production data - so don't do anything (deleting records, for example) here that you wouldn't want to be doing to your non-alpha Accelo account.
There are a few caveats with respect to this new alpha platform that you should know before diving on in:
The final point to make about this new infrastructure is that we're using some new (for us) cloud techniques to automatically scale up and scale down application servers to reflect load and demand - so as things heat up on Monday mornings more servers will automatically come online and as things quieten down on a Friday night servers will be stood down automatically too. We'll be watching logs, load levels and tuning the auto-scaling process accordingly - so if you see any periods where things seem more laggy than you think they should be, please let us know too (even though it isn't a bug per-se).