What is SigSentry?
AI-powered incident diagnosis that reads your logs, correlates code changes, and delivers structured root-cause findings automatically
SigSentry is an AI-powered incident diagnosis platform for engineering teams. You point it at your logs and code repositories, give it a short description of an issue, and it returns a structured diagnosis: severity, root cause, affected services, suggested actions, and the likely offending commit or file.
The problem we solve
When something breaks in production, the typical workflow looks like this. A support ticket or alert lands in someone's queue. An engineer opens CloudWatch or Datadog or Loki and starts grepping. They read 50 to 500 log lines, build a mental model, cross-reference recent deploys, and try to land on a root cause. Then they write up findings, tag the right team, and start the fix.
Most of that work — reading logs, correlating with deploys, drafting a writeup — is mechanical. It takes 15 minutes to several hours per incident even for experienced engineers, and it scales linearly with team headcount.
SigSentry takes the mechanical part off the table. The engineer still makes the judgment call on the fix; we just hand them a structured diagnosis to start from.
How we're different
We don't replace your log aggregator. We sit on top of it.
Your logs continue to flow into CloudWatch, Datadog, Loki, Splunk, Elastic, or GCP Cloud Logging — wherever they already go. SigSentry reads them on demand, scoped to the time window of an incident, and produces the diagnosis using an AI model with read-only access to your source code.
| Tool | What it does | What SigSentry adds |
|---|---|---|
| Log aggregator (CloudWatch, Datadog, Loki) | Stores and indexes logs | Reads them with intent and produces a diagnosis |
| APM (Datadog APM, New Relic) | Service-level metrics and traces | Identifies which deploy or PR likely caused an error |
| Incident management (PagerDuty, Opsgenie) | Routing and on-call schedules | Generates the postmortem after resolution |
| ChatOps (manual Slack queries) | Manual review of events | Triggered analysis from chat with /sigsentry analyze ... |
Who it's for
We designed SigSentry around three personas that share the same daily pain.
Support and on-call engineers drown in repetitive triage. SigSentry turns every analysis into a 30-second task instead of a 15-minute one, so the team can scale without hiring more rotations.
Engineering managers want consistent incident write-ups without pulling senior engineers into every incident. The auto-generated postmortem gives them that.
Customer-facing support teams typically can't triage tickets with deep platform knowledge. Connect Zendesk, Freshdesk, or Intercom and every inbound ticket gets a severity classification automatically — and the highest-severity ones can auto-trigger a full diagnosis.
What you'll get out of it
A SigSentry analysis returns a structured object you can read in 30 seconds: a one-paragraph summary, severity with confidence score, the root cause (service, error type, category), affected services and their roles, a reconstructed timeline, prioritized suggested actions, and — when a repo is connected — code correlation pointing at the file and pull request likely responsible.
You can generate a Markdown postmortem from any analysis with one click, ask follow-up questions in a thread, and rate accuracy to feed your team's quality metrics over time.
Ready to see it? The next page covers how it works end to end, or you can jump straight to the quick start to get a real diagnosis on your screen in five minutes.