|Database Server||Data centers|
|Web sites, API, webservices||Data centers|
|Documents and images||Azure|
|SMS||via Bulk SMS - http://www.bulksms.com/|
|Authentication||Auth0 - https://auth0.com/|
We use two data centers at geographically separate locations provided by Stratogen:
The web sites are load balanced over the two data centers. Failure at one data center will result in all web traffic automatically being routed to the other.
The databases are mirrored between the two data centers. If the secondary data center fails the primary will continue to operate as normal. If the primary fails we operate a warm fail over, whereby a minor manual configuration change is required.
All of our timesheet images and documents are stored in Azure.
We use European data centers with Geo Redundant Storage (GRS), which are located in:
All database backups are stored on a distributed file system, on the SAN replicated between the Data Centers.
The database transaction logs are backed up every 15 minutes. If both sites go down, at exactly the same time, the maximum data loss would be 15 minutes.
A full database backup is done daily.
We have offices in the UK and Australia, where the operations team monitor the site 24*7.
Continual monitoring of the system is done through using a variety of tools including:
We also have automated tests that run continuously, from three locations: Chicago, Dublin and Sydney. These tests perform basic availability and responsiveness checks and notify us immediately if the web site becomes unavailable from any of these locations.
Any and all alerts are e-mailed to dev and ops as well as being available to development and operations teams for further analysis.
Backups are tested daily through being restored to our test environment.
In the unlikely event that we cease trading, we will provide, as a minimum, one month’s access to the system, with the aim of providing three months.