There’s no shortage of tools for monitoring containers these days, and an increasing number of commercial offerings are incorporating the open source Prometheus monitoring software managed by the Cloud Native Computing Foundation (CNCF) to extend their capabilities.
One of the latest such initiatives comes from InfluxData. The latest release of the company’s namesake monitoring software supports the Prometheus standard exposition format via integration with Prometheus service discovery and scraping into Kapacitor, a real-time streaming data processing engine and metric read/write application programming interface (API) interoperability with InfluxDB, the InfluxData Time Series database.
Mark Herring, chief marketing officer, says InfluxData is not trying to be all things to container monitoring. Rather, the company is specifically focused on use cases where DevOps teams need insight into containers in real time. To facilitate that goal, the latest release provides extended support for the Kubernetes container orchestration engine. In addition, to employ InfluxData to Kubernetes to autoscale a cluster based on CPU or memory metrics, DevOps teams now also can scale clusters based on metrics such as queue depth or requests per second.
Herring says the number of instances where DevOps teams need access to real-time data analytics is increasing, especially in environments where there are many ephemeral containers that might run only for a few minutes. In addition, internet of things (IoT) applications that employ containers make extensive use of time-stamped data. To further that goal, InfluxData has joined the Eclipse Foundation’s IoT Working Group.
Having access to time-stamped data substantially reduces the amount of time and effort required to determine the source of any specific issue, he says. Today many DevOps teams spend more time in “war rooms” determining the source of problem than it takes to remediate the problem.
Herring says InfluxData plans to provide hooks into its database and processing engine to make it feasible to apply machine and deep learning algorithms against all the data collected about an IT environment.
Given the level of support for Prometheus, it’s clear there is now what amounts to shared research and development for container monitoring. Over time, various Prometheus technologies will be incorporated into container monitoring platforms that focus on areas that Prometheus does not address or that make it easier to analyze data generated by Prometheus.
It’s too early to say how much more broadly monitoring tools will be applied to containers. Various types of monitoring tools have been around for years, but the cost of employing those tools has limited their usage to mission-critical applications. However, given the complexity of microservices based on containers, many IT organizations will need to rely a lot more on monitoring tools just to keep track of what container went where, let alone what impact it’s having on the overall environment.