Industry executives and experts share their predictions for 2019.  Read them in this 11th annual VMblog.com series exclusive.

Contributed by Navdeep Sidhu, VP of Product Marketing, InfluxData

Monitoring Kubernetes in 2019 Will Get A Whole Lot Easier

In 2018, we saw a meteoric rise in the use of containers and orchestration technologies, such as Kubernetes. As workloads are shifted on to containers, DevOps teams are learning that legacy monitoring tools are not suitable for this new deployment architecture. From lack of granular visibility to the inability to handle ephemeral workloads, there are dozens of reasons why legacy monitoring is not able to keep up.

In 2019, we predict that the move to containerization and Kubernetes will gain further ground. As Kubernetes matures and gets further adoption, we expect to see a number of trends in the new year:

1. Evolution of Metrics and Events Observability: Growing deployments of apps on Kubernetes will require advanced monitoring technologies to support both metrics and events. Prometheus users will seek to complement Prometheus' pull-based model for scraping metrics from endpoints with other means to handle push-based events. Pull-based infrastructure metrics will be contextualized by push-based events to deliver observability.
2. CX Monitoring: Metrics and events will grow beyond monitoring infrastructure and DevOps toolchains. The need to monitor customer behavior and usage patterns will increase the need to monitor user metrics, such as clickstreams, and weave them into infrastructure metrics to drive "aha moments," like "my system usage went up 50 percent with a 30 percent increase in placed orders."
3. Time Series Platforms: As enterprise needs drive DevOps teams to look beyond the capabilities of Prometheus, time series platforms will gain more traction. DevOps teams will look at either complementing Prometheus with time series platforms or replacing it altogether to gain high-availability, fine-grained security and alerting.
4. The Rise of Purpose Built: Organizations will increasingly look to purpose-built time series platforms due to their ability to solve problems that require specific storage, handling, querying, compression, etc. Band-aid solutions, such as adding another Big Data platform to store time series data, are not sufficient.
5. Kubernetes will not eat the world: Kubernetes will gain ground, but there will still be workloads that will continue to operate on non-K8s infrastructure. Monitoring teams will adopt time series platforms that provide visibility into both K8s and non-K8s infrastructure and deliver a single pane of glass for monitoring infrastructure, DevOps, IoT, etc.

##

Navdeep Sidhu, VP of Product Marketing, InfluxData