Announcing free infrastructure metrics + automatic deployment and error analysis

Talia Moyal

by Talia Moyal

Explore more Announcements Blogs

Talia Moyal

by Talia Moyal


Looking for Something?

No results for 'undefined'

We’re excited to announce major updates to Lightstep’s observability platform.

Today we launched support for infrastructure metrics (!), automated, rapid error analysis, and version comparisons of canaries and other deployments.

With this release, Lightstep becomes the quickest way to understand what’s changed in your system and services.

A few key takeaways:

  • Telemetry data — including infra metrics — should be a commodity, and we’re leading the way. Lightstep customers (hello there) get access to memory, CPU, and network metrics, for free. This is in addition to end-to-end tracing, unlimited data, and system insights like Correlations.
  • Our new Error Analysis intelligence can resolve complex, multi-service regressions in less than 60 seconds. Yes, you read that correctly. If you want to see for yourself, check out our interactive sandbox.
  • You can compare deployment versions instantly. Simply deploy and we’ll automatically show you what’s changed in the new version.

Our goal is to enable developers to connect the dots from a regression at the top of their stack – even in a mobile or web client – all the way down into cpu, memory, and garbage collection metrics within their infrastructure.

A Better Way to Understand Infrastructure Metrics

We can analyze metrics the same way we do your trace data: by surfacing exactly what you need to resolve an issue or improve performance. You’ll be able to see memory, CPU, and network (as well as additional metrics specific to Go, Java, Python and NodeJS), alongside latency, errors and throughput, all in the context of your deployments, eliminating the need to flip between dashboards and tabs.

Lightstep metrics

We’ll also automatically highlight any changes to your service’s metrics that correlate with the regressions you are investigating. And since we also know your upstream and downstream dependencies, we can correlate changes anywhere else in your call stack.

Never Be Surprised by a Metrics Bill

You can instrument freely, even with high-cardinality tags, and remove any concern about what it will do to your bill.

This eliminates concerns of cost and cardinality, enabling you to get the insights you need without any of the stress around how it might affect your budget.

Automatic Error and Latency Analysis

We’ve made it easy to identify where errors are occurring and pinpoint which tags were the most correlated to those errors.

rca diagram 1

This means that in less than a minute, you can go from observing an error rate regression to identifying precisely where errors are originating from. Lightstep also shows you a prioritized list of which tags are most correlated with errors.

With latency analysis, you can clearly see which tags were most correlated to latency, as well as which operation contributes the most to both latency and the critical path.

rca latency tags 1

Monitoring Canary Deployments

With our Service Health feature, you can automatically track deployments as they rollout, in real-time, and how it’s impacting your SLIs.

You can see the independent impact of your release and then dive deep into latency or errors to understand what happened.


All of these features work seamlessly with our automatic deployment detection, meaning you can see before-and-after views to understand exactly how your system and services change post-deployments.

comparison analysis

Automated Insights: Our Deployment Report Card

We bundle this all into Automated Insights, which will surface recent deployments and regressions that pertain to services you might care about — automatically curated and shared with your team.

Improved Trace View

We’ve redesigned our Trace View page for easier navigation and increased visibility within a trace. The new design simplifies analysis along the critical path, and allows you to find problematic spans by viewing log messages alongside each span and by drilling into tags and span metadata. You can also jump into relevant parts of external tools, such as Github, from within the Trace View. The revamped interface also allows for the easy transition from a single span to a live, aggregated query.

logs in traces

How Can I Try Out These New Features?

Glad you asked. Options abound!

Interested in joining our team? See our open positions here.

Explore more Announcements Blogs