APM Solutions Venture into AI and DevOps
For a market segment as stable as application performance monitoring (APM), for individual vendors—even mature established vendors—to stand out requires a unique approach or feature set. At the very least, APM tools must now keep tabs on applications running in myriad locations, whether on-premises, on public or private cloud platforms, distributed as a microservice, or on a mobile device.
In his latest report, “GigaOm Radar for Application Performance Monitoring (APM),” analyst Ron Williams evaluates a number of current APM offerings. He reports standards within the APM space are minimal, but several solutions are starting to differentiate themselves by sharing data in vendor-agnostic formats.
The use of AI is increasing among virtually all vendors reviewed in this report. While AI/ML can analyze massive amounts of data and more quickly provide valuable insights, all the APM vendors are delivering differing levels of true AI functionality.
Adopting OpenTelemetry is becoming a factor, but remains slightly less prevalent than AI. The open-source project OpenTelemetry could provide a consistent method of transferring vendor-agnostic libraries and APIs to help gather and distribute trace, metric, and log data; which should help APM vendors focus more on system functionality and less on data transfer. While almost all vendors are incorporating AI, not all are embracing OpenTelemetry, so this will remain a factor to watch.
Table 1: Impact of Features on Metrics
Some of the more established APM vendors have made implementation easier within existing environments already invested in their other solutions. This includes companies like BMC, Broadcom, and Micro Focus. Adopting their APM offerings is more straightforward when added to their other infrastructure components.
Not surprisingly, Microsoft follows a similar path. In his report, Williams states Microsoft’s APM tool fits in well for shops adopting “everything Microsoft.” There are advantages to this approach, as other APM solutions can indeed monitor .NET environments, but don’t quite reach the scope or depth of which Microsoft is capable.
Among the more present challenges to be mindful of when considering APM are deployment and implementation costs. These costs can be a challenge to accurately determine across hybrid cloud and on-site environments. Vendors naturally provide professional services to assist with implementation, but that further adds to the cost.
Something to watch for in the future of APM solutions is increasing levels of DevOps integration and predictive analysis. IBM Instana is already moving in that direction by featuring CI/CD integration, and the Netreo now features anomaly prediction.
Read more here: Key Criteria for Evaluating APMs, and Gigaom Radar for APMs