"[Struct is like] a junior engineer with infinite energy and a genuine desire to go investigate whatever it finds."

Sri Raghavan

Founding Engineer, Orb

About Orb

Orb is the billing platform for modern software companies, powering usage-based pricing, metering, invoicing, and revenue workflows.

The Challenge

Orb's engineering team was fielding a steady stream of production signals across a sprawling toolchain - Sentry exceptions routed into Asana, monitors firing into a noisy #prod channel, customer-reported issues arriving through support intake, and incidents spun up in Incident.io.

High-priority Sentry alerts required manual triage that pulled on-call engineers across Datadog, CloudWatch, GitHub, and Slack threads to reconstruct context for every issue, and the noisiest surfaces were getting muted rather than investigated.

As Sri put it, the manual process was "error-prone, especially for junior engineers," and hiring more engineers to handle the growing alert volume wasn't a scalable solution.

The Solution

Orb connected Struct to Sentry, Asana, Slack, GitHub, Datadog, and CloudWatch, starting with a single scoped surface: auto-investigating every Sentry alert that hit their Asana queue. When an alert fires, Struct investigates automatically - querying logs, analyzing stack traces, reviewing recent commits - and posts a full investigation with root cause and supporting evidence before an engineer picks it up.

"People are pleasantly surprised by how often Struct gets the explanation right." — Sri

As confidence grew, Orb expanded Struct organically to three more surfaces:

• Bugs & FDE intake — auto-investigating customer-reported issues packaged by technical support engineers

• On-call incidents — Struct joins every Incident.io channel as it opens, pulls the originating context, and posts an initial investigation

• Product on-call — investigating issues routed to the product on-call queue

"[We were] doing Sentry investigations. Now we're automatically doing bug investigations, and incident investigations. I think that just adds up." — Sri

The Results

Usage grew organically and doubled in two months. In the last 30 days, Orb increased usage +34% and expanded from one to all new incident channels.

That trust showed up early. Days after Struct went live on their Sentry queue, Orb ran a database failover that scattered connection errors across services:

"We did a DB failover…(so, connection drops across lots of different services), and the investigations are on point." — Sri

Adoption spread organically. Engineers beyond the on-call rotation began pulling Struct into their own workflows - bulk-assigning invoice-finalization alerts for investigation, requesting auto-PRs for fixes Struct had already found, and asking to point Struct at new Asana queues.

Investigations compound. Struct remembers prior investigations and feedback, so its answers improve with use. When a second customer reported a data-export delay a week after a similar real incident, Struct compared the two and flagged that this time, the metrics showed it wasn't lag at all:

"Struct was like: unlike this other investigation I did last week and this other incident that we had last week, this is actually not lag... it doesn't have amnesia. It remembered something."

Sri Raghavan

Fouding Engineer, Orb

A close working partnership. Orb and Struct iterate together, from tuning noisy alert channels to designing incident-channel workflows and on-call handoff summaries, with feedback from Orb's on-call rotations shipping as product improvements within days.

Conservatively, at ~15 minutes saved per investigation & incident, Struct gives Orb's engineers 160 hours/month back — a >10x return on their investment — while coverage keeps expanding across every surface where production issues surface.