Our Servers

Your Atlassian application instances run on:

Supermicro servers with 2 x (Intel Xeon E5-2630 v2 @ 2.60GHz), 256GB of RAM, and 2 x 512GB SSD deployed in Q4 2014.

A few are on HP ProLiant 360p G8 with 192 GB RAM and 8x HDD RAID-10 deployed in Q4 2012.

We continuously migrate old instances to new servers as part of our promise to always pass on the benefits of any upgrades to all our customers rather than only to new ones.

Data Centers

Our servers are co-located in long-established, Tier-3 data centers, compliant with ANSI/TIA-942. ndc_2_m

The highest level of service is guaranteed by:

guaranteed uninterruptible power system (UPS and power generators)
redundant cooling system
data center connected to a fiber optic ring with two separate routes
redundancy of network devices
dedicated vlan (1Mbit-10Gbit) and the lease of dark fiber to nodes of other operators
the entire infrastructure is based on the best solution providers, such as: EATON, Moeller, APC, Hewlett-Packard
raised floor based on steel profiles with a lifting capacity 1500 kg/m2
hot/cold corridor system
security and monitoring facility
fire protection system
servers rooms built with thick walls without windows
servers rooms are 4 meters above ground level (no flooding)
buildings are located on hills, not in the floodplain
no additional floors in the building above the Data Centers
there are no potentially dangerous objects in the immediate area, such as gas stations

Networking

A modern high-speed network, the network used by our data centres has the topology of many rings. In case of equipment failure or interruption of any connection (eg. broken optic fiber) traffic is automatically switched to the alternative route.

Simplified diagram of one network structure:

With excellent connections with other networks, open and proactive peering policy and peering with operators from neighbouring countries.

Service Level Agreement

Check our full Terms & Conditions

Despite our basic plans don’t come with guaranteed SLA we strive for the best nontheless:

our servers didn’t have any unplanned outages in 2013 and 2014
there were only 2 planned maintenance downtimes, 4 hours long each, which resulted in 99.9 availability.

24/7 automated monitoring and alerting systems (on separate servers of course) keep an eye on your application instances. Support is notified whenever a failure of any instance is detected.

Checks are performed once a minute and the average availability of all the applications at the moment is:

99.78546 for last 7 days
99.94957 for last 30 days
99.83620 for last 365 days

As of 2014 any downtime, being it planned or unplanned, is included in the numbers above.

Backup policy

We perform daily backups of files and databases. Backups are sent to a separate physical server in a physically different data center and they are kept for one year.

Our customers are able to download backup files via SCP or request us to upload them where more convenient for them.

Maintenance policy

There are no more than four maintenance downtimes a year and on average only 2/year. They are planned up-front, performed during the Saturday nights (Europe time) and last no more than 4 hours.

We give our customers considerable notice about any planned downtime via both direct email and social channels.