Running in production

Using analysis tools in development is one thing; running them in production is another. Running a single container on one machine is not an acceptable strategy for operating a service that provides information that's critical to an organization. It's worth considering the challenges of scaling telemetry backends to meet the demands of the real world. The following subsections highlight areas that require further reading before you run any of the backends mentioned earlier in production.

High availability

The availability of telemetry backends is likely not as critical to end users as that of the applications they are used to monitor. However, having an outage and realizing that the data that's required to investigate is unavailable or missing during the outage causes problems. If an application promises an uptime of 99.99%, the telemetry backend must be available to account for those guarantees. Some aspects to consider when thinking of the...