As IBM continues its transition from a software and services company to a hybrid cloud management company they have made a number of investments to facilitate this transition. Most recently, IBM announced their acquisition of Instana, an application performance management (APM) startup with a cloud-native approach. The deal is expected to close in the next few months and is subject to regulatory approvals; neither company has disclosed a purchase price for the deal.
Why This Transaction?
The York IE team chose this as our transaction of the week to discuss IBM’s transition to a hybrid cloud management company. Prior to their latest acquisition, IBM provided a portfolio of Tivoli tools for their APM; now, they can integrate Instana’s APM and add their own artificial intelligence to advance their AIOps and IT automation efforts. In the past, IT teams only used APMs to monitor mission-critical apps due to the cost and complexity related to deploying your own APM. Now, as APMs are easier to deploy, more teams are running them; at the same time, applications nowadays require consistent vertical scaling and experience far more traffic. Because of this, sophisticated APMs are now including observability tooling to beef up their sampling frequency and capabilities. Originally, APMs ran samplings roughly twice a minute and would report issues to admins for them to go in and fix. With observability tooling built into an APM, the sampling process becomes a “gauntlet” for applications with samplings occurring every second. High-frequency sampling is a must-have for applications that process large numbers of transactions or even applications that frequently roll out updates.
Instana’s APM: Without Samples
A unique feature that sets Instana’s APM apart from other competitors is the fact that their APM does not employ sampling but rather uses tracing. Instana’s SaaS backend analyses more than one million traces per second whereas other observability platforms are often just sampling their traces. Two major flaws in only sampling traces are that more often than not, normal traces will be dropped to free up storage and many platforms have maximum throughput limits. With their instrumentation code, Instana’s APM runs every trace without any limits creating the most complete picture and providing DevOps engineers with data for future optimizations.