Hermetiq / Build Analytics & AI

Build Analytics & AI.

With Barn Owl, your AI build engineer.

Barn Owl is Hermetiq's AI Agent that optimizes your Bazel and Buildbarn builds. Keep your existing infrastructure, or use ours.

You can use Barn Owl directly in Hermetiq, or give your own AI agent access to Barn Owl via MCP.

Try Hermetiq Cloud →Try Hermetiq on-prem →

Discuss your deployment →

From symptoms to answers

Barn Owl helps you:

Find and fix build bottlenecks—even with remote execution
Pinpoint what changed when builds get slower
Understand cache misses and stop rebuilding the same work
Find out why builds are queueing and workers are idle
Diagnose intermittent remote build failures
Improve your Buildbarn configuration with reviewable changes
Identify wasted infrastructure capacity and reduce build costs

Investigate with context. Act with evidence.

Barn Owl brings together your build data, infrastructure telemetry, and configuration to help explain what’s happening and recommend what to do next.

Bring us your toughest build problem →

Autonomy you grant—not autonomy it takes

Earns its place in your stack.

It reads the build graph before its first coffee. Because it doesn’t drink coffee. Because it’s an owl.

Wide awake.

Cache hit-rate erosion. A widening critical path. A worker pool quietly burning budget. Barn Owl watches the signals in your stack around the clock.

Knows your build.

Fluent in Bazel and Buildbarn. It connects actions, cache events, queues, workers, and configuration to explain what changed and why.

Earns your trust.

Explanations first. Recommendations next. Then fixes as PRs you review and merge. More autonomy only where you grant it—never the keys to prod by default.

Gets wiser.

Every build outcome is another lesson in your stack’s failure modes. The owl learns what actually worked, not just what sounded plausible.

Build expertise. Context included.

Barn Owl combines Hermetiq’s tools and data with purpose-built prompts, model routing, and context from your builds, remote execution environment, and configuration. It brings that expertise into the part of Hermetiq where you’re working.

AI diagnosis of build failures and cache misses
Queue, worker and storage investigation
Reviewable build and infrastructure fixes
Build performance and capacity optimization

Connected evidence

The data behind your build analytics.

Connect what happened in a build to the actions, cache, and infrastructure behind it.

Bazel build data

Build Event Protocol (BEP)
Builds, targets, tests, failures, and configuration events.
JSON trace profiles
Build phases, action timing, parallelism, and resource counters.

Buildbarn action and cache data

Action Cache lookup events
Cache hits and misses, linked to the actions behind them.
Completed-action logs
Remote execution timing, results, and action resource usage.
Content Addressable Storage (CAS)
Commands, inputs, environments, and platform details that explain cache behavior.

Infrastructure telemetry and costs

Metrics via VictoriaMetrics
Buildbarn and Kubernetes performance and health signals.
Logs and events via VictoriaLogs
Pod logs and Kubernetes events for operational diagnosis.
Kubernetes API
Worker pools, workload configuration, replicas, and autoscaling state.
OpenCost
Infrastructure costs and resource allocation data.

Available analytics depend on the data sources connected to your deployment.

See how each data source connects →

BB Config Editor + Barn Owl

The same Barn Owl, focused on your configuration.

Work with an agent that understands your open configuration, uses specialized Buildbarn prompts, and selects models based on your question.

BB Config Editor

Inspect, edit, render, and validate your Buildbarn configuration. Work with Git-backed ConfigSets, import and export ConfigMaps, and review validation errors alongside your source.

Barn Owl

Ask questions about that configuration, investigate problems, and review proposed changes. Connect telemetry so its recommendations can draw on your running environment.

Start with your configuration. Connect live telemetry for deeper diagnosis. Barn Owl can explain the files you load before telemetry is connected; understanding live worker, queue, and storage behavior requires the relevant signals from your deployment.

BB Config Editor showing storage.jsonnet, a deployment file tree, and Barn Owl explaining the content-addressable storage data integrity validation cache — Ask about the configuration you’re editing. Barn Owl explains it alongside the source.

Ask your config anything.

Get answers grounded in the ConfigSet you load, with the context to understand how your configuration fits together.

Fixes proven before you see them.

Barn Owl finds the file that broke your render and proposes a repair that already renders and validates. You review. You apply.

A storage audit that knows the pitfalls.

Block geometry, spare blocks, key-location maps, hardcoded namespaces. Every finding cites the file and path so you can see exactly what needs attention.

Recommendations backed by measurements.

Connect storage, scheduler, and worker telemetry to guide recommendations with measurements from your environment.

Diagnose idle workers in seconds.

A platform mismatch, a capacity ceiling, or an empty queue? Barn Owl tells you which so you can take the right next step.

Security rollouts, planned for you.

Plan OAuth2 and mutual TLS across every Buildbarn service, with validated proposals and a clear view of the ports that must stay open.

More than a wise face

The context to diagnose. The tools to fix.

Buildbarn Config Editor + AI Assistant

Stop hunting through repos for the right proto field. Ask an AI assistant that has indexed Buildbarn’s protobuf schemas, then edit configuration in forms generated from those same schemas.

Changes are saved as git commits, with optional PR review—so your team can see what changed, who changed it, and review it before deployment.

Your build context. In your coding agent.

Hermetiq’s MCP integration gives compatible coding agents access to your build evidence. Bring failures, cache misses, and execution context into the workflow where you already write code.

Explore MCP for Bazel →

Choose where it runs.

Use Hermetiq Cloud or deploy in your private cloud or data center.

Cloud

Analyze your Bazel build data with analytics and Barn Owl AI. Keep your existing infrastructure.

Who operates your infrastructure?: You keep and operate your existing build infrastructure
Where Hermetiq runs: Hermetiq Cloud

Connect Bazel build events to Hermetiq. Add available cache, action, and infrastructure telemetry from Buildbarn for deeper analysis.

Setup guide

SOC 2 Type I · Trust Center

On-prem

Analyze your Bazel build data with analytics and Barn Owl AI. Keep your existing infrastructure.

Who operates your infrastructure?: You keep and operate your existing build infrastructure
Where Hermetiq runs: Your private cloud or data center

Connect Bazel build events to Hermetiq. Add available cache, action, and infrastructure telemetry from Buildbarn for deeper analysis.

Setup guide

Build Analytics & AI is a complete offering and is also included with Managed Cache & Execution. Barn Owl is the AI build engineer inside Hermetiq. It watches your connected telemetry, explains what changed, and proposes fixes as PRs you review and merge. It learns from build outcomes and earns more autonomy only where you grant it. Hermetiq’s MCP integration also brings build context into your existing coding agents.

Plan your deployment.

Talk through your platforms, infrastructure and build goals with us.

Contact Jake for pricing (Jake@Hermetiq.com)

Discuss your setup Read the Quickstart

Move from an infrastructure symptom to the workload behind it.

Buildbarn exposes powerful infrastructure primitives. Hermetiq connects those signals to the Bazel builds, targets, and actions your team is trying to improve.

Queue pressure

Which platform queue is saturated, and which builds are creating the demand?

Action latency

Are slow actions waiting for a worker, transferring inputs, or executing?

Cache behavior

Did cache behavior change because action inputs, commands, platforms, or environments changed?

Cost attribution

Which builds, targets, and action types are driving storage and execution cost?

Connect your build data. Keep your infrastructure.

Connect build events and the action, cache, worker, queue, storage and configuration data available from your deployment so the agent can trace problems to their causes.

Connect Bazel build events

Send structured Build Event Protocol data so Hermetiq can understand invocations, targets, tests, timing, and outcomes.

Connect available Buildbarn signals

Add the completed-action, metrics, logs, or traces exposed by your deployment and observability stack.

Let the owl earn its place

Barn Owl starts by explaining and recommending. Connect your configuration repository so it can propose fixes as PRs you review and merge. Grant more autonomy only where it has earned your trust.

See the operational layers separately—and together.

Remote execution phase analysis

Separate queueing, input transfer, execution, and output transfer instead of treating remote time as one number.

Build and action cost attribution

Connect infrastructure consumption to builds, targets, and action types so teams can prioritize the expensive work.

Let the owl take the night shift.

Connect your deployment to Hermetiq Cloud, or install intelligence in your own environment.