disaster recovery plan

X-Guard BV has a description, contact persons and plan per service where a malfunction or incident can occur. All elements in our services are part of a ''self healing auto scaling microservice architecture''. In concrete terms, this means that in almost all cases a problem is solved by smart horizontal scaling. Of course, this does not mean that we cannot prevent problems that require more than auto-recovery.

The plans per service are not public. These contain personal data of staff and details about network and software architecture that are not shared publicly (or with customers). In our SLA defines the obligation we must meet to guarantee the continuity of the service. Continuity is very important: we work with alarms and any alarm that cannot be sent for whatever reason is a serious situation.

The X-Guard IT service desk is 24/7 available to solve any problems. As mentioned in the SLA, all communication about malfunctions and maintenance is conducted via https://status.x-guard.nl.

All elements in our microservice architecture are automatically tested (weekly) via a tooling that crashes 'random' elements. This happens, for example, with tooling such as Pumba (https://github.com/alexei-led/pumba). This ensures continuity and answers the question of when testing was last performed.

Almost all updates we post are ''zero impact'' and are rolled out during office hours at the beginning of the week. If an update does have downtime, an announcement will be made no later than one week before the maintenance, including the time window. Determining the window is based on the content of the update. There are no generic guidelines for this because each update has specific points of interest that influence the decision for the right time.

Each version of the app, as well as other software components, is always extensively tested in several phases:

  • Local development
  • Pull request reviews
  • Shared remote development testing
  • Alpha testing
  • Beta testing
  • (Release candidate)
  • Stable

In all phases of the development process, close attention is paid to ensure that no new problems are introduced. The application is actively tested on ±10 different phones and periodically tested on another ±20 phones.

Looking for more information?