Splunk pricing worries some users amid SignalFx monitoring buy
Splunk wants to collect more data via SignalFx, but some users wonder how much it will cost to keep up with the proliferation of logs and metrics in the age of cloud monitoring.
Splunk broke open its piggy bank and shelled out $1.05 billion for cloud monitoring vendor SignalFx this week. The move reflects a broadening scope for tools to monitor complex IT environments, but some users worry about overreach into their IT budgets.
Splunk, which began as a log analytics player, will gain expertise and software that collects metrics and distributed tracing data in cloud-native application environments with this acquisition, as it sets its sights on speedier data analytics features that support a broader swath of data sources.
The deal, valued at $1.05 billion, is the largest in recent memory in the IT monitoring space, analysts say. It also indicates future trends that will see cloud monitoring tools for increasingly complex applications and software frameworks such as container orchestration and service mesh take over enterprise IT. It's the third merger between IT monitoring and automation firms in the past week -- IT infrastructure monitoring firm Virtual Instruments also acquired cloud monitoring startup Metricly, and IT automation vendor Resolve Systems bought AIOps player FixStream.
"There's a lot more interest in monitoring these days," said Nancy Gohring, analyst at 451 Research in New York. "I would tie that to the growing interest in cloud and cloud-native technologies, and the maturity curve there, especially from the enterprise perspective."
Splunk pricing a potential downside of cloud monitoring updates
[Enterprise IT] shops are learning the hard way that you have to change your approach to monitoring in these new environments, and, thus, there's more demand for these types of tools.
Nancy GohringAnalyst, 451 Research
Splunk users are aware of emerging trends toward cloud-native application monitoring, streaming data analytics and machine learning, and see the need for fresh approaches to data collection and data analytics as they grow.
"It lowers the entry barrier for machine learning and gets people asking the right questions," said Steve Koelpin, lead Splunk engineer for a Fortune 1000 company in the Midwest, of Splunk's Machine Learning Toolkit. "People share their code, and then others use that as a starting point to innovate."
But an emphasis on machine learning that includes metrics data will demand large data sets and large amounts of processing power, and that comes at a cost.
"It's expensive to stream metric data as it's metered at a fixed 150 bytes," Koelpin said. "This could mean substantially [higher] license costs compared to streaming non-metric data."
Splunk pricing starts at $150 per gigabyte, per day, with volume discounts for larger amounts of data. The costs for Splunk software licenses and data storage have already driven some users away from the vendor's tools and toward open source cloud monitoring software such as the Elastic Stack.
Tim Tully
"We're always evaluating pricing options, but our focus is more on making sure we build the best products we can," said Tully, senior vice president and CTO at Splunk, when asked about users' Splunk pricing complaints in an interview last month.
Splunk offers the most extensive set of log analytics features on the market, as well as good data collection and analytics performance and stability, Gohring said. However, some users chafe at paying for that full set of features when they may not use them all.
"People want to collect more and more data, and there's always a cost associated with that," she said. "It's something all vendors are struggling with and trying to address."
Splunk prepares to gobble up massive data
As enterprises adopt cloud-native technologies, they will look to vendors such as Splunk, pricing notwithstanding, for cloud monitoring tools rather than build the cheaper homegrown systems early adopters favored, Gohring said.
"Those shops are learning the hard way that you have to change your approach to monitoring in these new environments, and, thus, there's more demand for these types of tools," she said.
In fact, 451 Research estimated that container monitoring tools will overtake the overall market revenue share held by container orchestration and management tools over the next five years. A June 2019 market monitor report on containers by the research group estimated total application containers market size at $1.59 billion in 2018, and projected growth to $5.54 billion in 2023. In 2018, management and orchestration software vendors generated 32% of that revenue, and monitoring and logging vendors did 24%. By 2023, however, 451 Research expects management and orchestration vendors to generate 25% of overall market revenue, and monitoring and logging vendors will do 31%.
In the meantime, Splunk plans to capture its slice of that pie, when beta products such as its Machine Learning Toolkit, Data Fabric Search and Data Stream Processor become generally available at its annual user conference this October. These products will boost the data collection and query performance of the Splunk Enterprise platform as it absorbs more data, and give users a framework to create their own machine learning algorithms against those data repositories under the Splunk UI. Splunk will also create a separate Kafka-like product for machine data processing based on open source software.
"We're looking to add Apache Flink stream data processing under the Splunk UI for real-time monitoring and data enrichment, where Splunk acts as the query engine and storage layer for that data [under Data Fabric Search]," Tully said.
SignalFx has strong streaming analytics IP that will bolster those efforts in metrics and distributed tracing environments, Gohring said.
Tim Tully
