This guide provides best practices, as well as important considerations for operating production systems. It covers the following topics:
- Device access and management
- Network access
For operating production deployments, we strongly recommend using software to manage your fleet of physical or virtual devices. We will refer to this as Unified Endpoint Management (UEM). Depending on where you look and the exact use case, such software may also be referred to as Mobile Device Management (MDM) or Enterprise Mobility Management (EMM). UEM systems offer a lot of features, such as remotely accessing or configuring devices. In the following sections, you can find an overview of the operational aspects that cause this recommendation.
In a production environment, there are several situations that will require you to update your app(s), or Actyx. As most solutions run on several devices, physically going to one after another to do updates is not a viable solution. With UEM systems, you can roll out updates of your app(s) or Actyx remotely, and automatically.
If things go wrong, you need a way to access the logs of your app, logs of Actyx, and logs of the device host system. Without a proper setup, you again have to go to a device and wire your laptop to it to get these logs. A UEM allows you to remotely collect device logs.
UEM systems allow you to remotely configure devices and standardize device configurations across your fleets. This helps you when adding or removing devices and facilitates standardization (e.g. network configurations).
With regard to host configuration, it is particularly important to configure device clocks correctly. We strongly recommend using NTP and also monitoring the clocks for deviations, since misconfiguration or user error may result in the clock being off by weeks (the largest we observed so far was that a tablet clock was set back by a year, which of course caused problems).
With a UEM, you can also enforce configurations and make sure the device clocks are not changed. Bugs, or rather unexpected app behavior has often been traced back to incorrect device clocks leading to misinterpretations of event timestamps. Also, when interpreting any logs (app, Actyx or host) from your device, incorrect clocks will lead to problems.
We have experienced production incidents that were caused by network or firewall changes that did not comply with Actyx network requirements. As MDM systems allow you to remotely access devices, they allow you to remotely test network requirements (see a guide for testing network requirements here).
For practical advice on choosing a mobile device management system, check out our how-to guide.
In addition to using an MDM, it is also recommended to have remote access to the devices' network. You can either remotely access the network from your PC (e.g. via VPN), or have remote access to a PC in the device network. For configuring Actyx nodes, or for troubleshooting certain issues, you will then need to connect to nodes via the Actyx CLI / Node Manager.