Fotolia

Application dependency mapping expands in New Relic One

DevOps monitoring vendors rev their engines as enterprises take on more sophisticated IT monitoring practices to reach the next major phase of app modernization.

IT pros tackled faster application delivery and advanced infrastructure automation as the DevOps movement first took hold. Now, sophisticated application monitoring is about to become the next big phase of tech transformation.

With highly distributed and automated applications and infrastructures in place, DevOps teams must effectively measure their performance and make performance data easily accessible to guide continuous improvement. Enter DevOps monitoring tools such as New Relic One, released this week, which can accommodate these needs with application dependency mapping that spans distributed enterprise DevOps teams and publicly available support for granular AWS Lambda monitoring.

"Modern environments have more shared resources, so if you're restricted to only looking at your app and your resources, you might not understand that someone else's app might be causing problems for you," said Nancy Gohring, analyst at 451 Research.

Cox Automotive goes for AWS Lambda monitoring update

New Relic previously offered application dependency mapping within individual accounts, but New Relic One establishes correlations, called Service Maps, between IT services that aren't managed by the same teams.

At Dealertrack Inc., a provider of finance and insurance software for the automotive industry, this new feature saved time for development teams to troubleshoot applications during beta tests of New Relic One.

If something wasn't working, we had to log into the console and look at logs in up to four or five different places. ... We've set up alerting now for single errors, and we're looking into anomaly detection for Lambda duration.
Matthew Vaughan Lead software engineer, Dealertrack

"We have three dashboards right now to monitor production and non-production accounts for our [AWS-based] replication pipeline," said Matthew Vaughan, lead software engineer at Dealertrack, a subsidiary of Cox Automotive Inc. headquartered in Lake Success, N.Y. "In New Relic One they all show up in the same list."

Vaughan's team built an event-driven data replication system that uses AWS Lambda, Fargate and other highly abstracted services, but doesn't manage all of those elements. New Relic One can pinpoint the root cause of issues with the replication pipeline even if Vaughan's team doesn't control the service involved.

"We depend on resources that other dev teams own, [such as] a Fargate instance we rely on to process data queues," he said. "Before, if something wasn't working, we had to log in to the console and look at logs in up to four or five different places, but we've set up alerting now for single errors, and we're looking into anomaly detection for Lambda duration."

The data replication pipeline supports a project at Dealertrack to modernize legacy applications based in IBM iSeries systems on premises and migrate them to the AWS cloud. This replication system primarily relies on custom Lambda functions and coordination between asynchronous services, which will require the features that come with updated AWS Lambda support in New Relic One.

"We don't get a lot of data on Redis instances by default on Amazon," Vaughan said. "New Relic One can drill down into that system and custom event logs go to the New Relic Insights time series database."

DevOps monitoring arms race intensifies

New Relic One also simplifies the query interface for IT pros not steeped in SQL. The New Relic One homepage UI features a query bar that can translate natural language into New Relic Query Language format. This brings it into closer competition with emerging IT monitoring tools such as Honeycomb, which focus on queries rather than graphs and visual dashboards to inspect monitoring data.

Cross-account application dependency mapping also isn't new to the industry. New Relic must compete with vendors such as ScienceLogic, Boundary, Dynatrace and SignalFx in addition to Honeycomb. On the serverless monitoring front is competition from specialists such as IOPipe, Thundra, Espagon and Dashbird, in addition to traditional rivals such as Datadog that offer serverless monitoring features.

All this market activity signals that enterprise IT is ready to refresh its approach to DevOps monitoring, Gohring said. Now that New Relic has overhauled its platform, it must quickly deliver more differentiated features based on its CoScale and SignifAI acquisitions.

CoScale and SignifAI engineers and IP played a role in New Relic One development, company officials said, and they plan to roll out more sophisticated log analytics and AIOps features in the second half of 2019.

New Relic One is available this week to all current Pro-level subscribers at no extra charge. For the next three to six months, AWS Lambda monitoring support is available without a separate license, though that may change by the end of 2019, company officials said.

Dealtertrack's Vaughan said he looks forward to AWS Lambda monitoring support for .NET-based functions, in addition to the Node and Python-based functions New Relic One supports with this release. He said he'd also like to see anomaly detection features available with New Relic's API so that he can invoke them with Terraform-based infrastructure as code.

Dig Deeper on IT systems management and monitoring