Skip to content

Startup’s eBPF APM instruments flip up warmth on Datadog

The DevOps group at a New York-based insurer will quickly place an enormous guess on a startup’s Kubernetes monitoring instruments, which may partially change Datadog due to an method to information gathering that requires no instrumentation.

Lemonade Inc., based in 2015, sells renter, house owner, automobile, pet and time period life insurance coverage via a digital platform that makes use of automation and analytics to maintain its labor prices low, even because it has grown to greater than 1 million paying clients. This method extends to the corporate’s DevOps group, primarily situated in Israel, which helps some 200 builders with 5 DevOps engineers.

Observability information is essential to driving the automation required to take care of this ratio whereas making certain resilient programs for about 100 utility providers, stated Avital Tamir, DevOps lead at Lemonade. Most of that information has been gathered via Datadog’s software program, which covers the IT infrastructure in addition to utility efficiency administration (APM).

The corporate grew during the last 5 years at a compound annual development fee of 150% and from 23 workers at its founding to greater than 1,000. So did the price of observability when it comes to software program licensing and developer time for guide instrumentation, Tamir stated.

“The billing for Datadog simply saved rising and rising. And searching a number of years forward, we form of figured we have to begin doing one thing that is a bit more economical,” Tamir stated. “Not solely that but in addition to open up the potential of storing metrics for lengthy durations of time … with excessive cardinality and with out going bankrupt.”

Tamir did not specify how a lot spending on Datadog had been projected to develop long run. However he did say its licensing prices had been anticipated to triple this yr in comparison with final yr. That put the purpose of storing extra information for longer to feed historic analytics engines largely out of attain.

Groundcover’s dashboard places eBPF information from Kubernetes into context and prioritizes alerts.

From drawings on napkins to APM instruments in manufacturing

Tamir joined Lemonade two years in the past. Just a few months into his tenure there, he heard about an early-stage firm referred to as Groundcover, primarily based in Tel Aviv. Groundcover’s founders had been engaged on utilizing a Linux kernel-based utility, Prolonged Berkeley Packet Filter (eBPF), to assemble utility efficiency information.

This isn’t distinctive. Demand for eBPF help to watch and handle Kubernetes purposes, from the community infrastructure the place open supply Cilium has made headway to service meshes and observability instruments, has been producing rising buzz within the trade since 2020. Current APM distributors equivalent to Datadog and New Relic may also collect information by way of eBPF and automate the instrumentation of apps for eBPF observability via an open supply utility named Pixie.

Simply the truth that there isn’t any question language it’s essential study — simply these good screens that take you thru the work of exhibiting you the place issues went unsuitable — could be very spectacular to me.

Avital TamirDevOps lead, Lemonade Inc.

Nonetheless, whereas Pixie makes instrumenting purposes simpler, what intrigued Tamir about Groundcover’s deliberate APM instruments was their promise to not require any instrumentation of Kubernetes purposes, automated or in any other case, to watch them.

As a substitute, Groundcover installs an agent by way of a Daemonset on Kubernetes clusters and makes use of info flowing out of the Kubernetes API to place that information in context for every utility. Groundcover’s agent additionally performs some analytics on the cluster degree earlier than sending information to its central repository. This reduces visitors on the community, trims the quantity of knowledge that have to be saved for long-term evaluation and prioritizes alerts in keeping with criticality.

Lemonade now makes use of Groundcover’s dashboards displayed on TV screens to indicate builders information about their apps in take a look at and improvement environments. Not having to instrument purposes individually for every programming language has allowed Lemonade’s DevOps group to roll it out, replace it and feed information to its packaged dashboards for these environments rapidly and simply to date, Tamir stated.

“Simply the truth that there isn’t any question language it’s essential study — simply these good screens that take you thru the work of exhibiting you the place issues went unsuitable — could be very spectacular to me,” he stated.

Groundcover, based in 2021, raised $24.5 million in seed and Collection A funding in September. Its main focus up to now is on Kubernetes purposes. Datadog covers the complete IT stack from utility to cloud infrastructure in addition to IT safety analytics.

Lemonade has but to make use of the software in manufacturing however plans to take action this yr, in keeping with Tamir.

“We have been watching [Groundcover] develop from drawings on napkins … and it feels prefer it’s simply getting began,” he stated. “I actually want to see it mature over the subsequent yr right into a product that may really rival a number of the massive observability gamers. However they nonetheless have some stuff they do not do but, like customized metrics or observing cloud sources like AWS.”

Nonetheless, if all goes properly with a manufacturing rollout this yr, Groundcover will start to exchange Datadog, beginning with APM, Tamir stated.

“We’re getting a head begin,” he stated. “Redesigning the stack is generally us trying 5 years forward. … And the important thing concept there’s not [for Groundcover] to be the most effective information analytics platform however a extremely nice troubleshooting platform that helps you floor issues and present you what went unsuitable earlier than you even know what to search for.”

Beth Pariseau, senior information author at TechTarget, is an award-winning veteran of IT journalism. She could be reached at [email protected] or on Twitter @PariseauTT.

Leave a Reply

Your email address will not be published. Required fields are marked *