Skip to content
Book a demo

Provenant

The operational intelligence layer for your engineering team.

Provenant investigates incidents the way your best engineer would — then remembers what your team confirms, so the next investigation starts ahead.

Capabilities

01

Evidence-backed investigation

Start from an alert or an operator request. Provenant selects allow-listed, read-only tools to query logs, metrics, queues, deployments and code — and keeps every result as cited evidence.

  1. Deploy a1f3c9e to payment-serviceDeploy
  2. Pool wait time p95 spikesPrometheus
  3. “Timeout acquiring connection” × 412Loki
  4. Settlement queue backlog growsRabbitMQ
Investigation timeline with tool calls and evidence references · Illustrative data

02

Code & change correlation

Runtime symptoms are traced to the deployed commit, the config and code that changed, and the services that depend on them.

deploy a1f3c9e → payment-service · 10:39

config/pool.yaml
-  max: 50
+  max: 10

error rate ↑ at 10:42 · 3 min after deploy

Deploy, commit and diff linked to an error spike · Illustrative data

03

Ranked hypotheses & RCA

Competing explanations are scored on the evidence. The final RCA separates facts from hypotheses and states its confidence.

  • Connection-pool size reduced in config changeSupported
  • Downstream bank API latencyWeak evidence
  • RabbitMQ consumer crashWeak evidence
Hypotheses ranked by supporting and contradicting evidence · Illustrative data

04

Institutional memory

Engineer-confirmed RCAs and resolutions are stored in a knowledge graph and surfaced automatically when similar incidents return.

  • INC-1873Pool exhaustion after config changeConfirmed RCA
  • INC-1650Settlement backlog during deployConfirmed RCA
  • RB-12Runbook: roll back payment-service configRunbook
Similar past incidents and runbooks · Illustrative data

Anatomy of an RCA

Every completed RCA answers the same eight questions.

Root cause
The confirmed cause with linked evidence
Confidence
Evidence-derived, with reasons
Blast radius
Services and dependencies affected
Use cases
Business flows impacted
Transactions
Deterministically counted
Customers
Unique affected customers, counted not guessed
Time window
Start, detection and recovery
Recovery
Verified against live signals

Coming next

One evidence foundation, more ways to use it.

  • Audit assistant

    Answer auditor questions with source-referenced incident, approval and provenance records.

  • Engineer trainer

    Service walkthroughs, historical incident replay and on-call readiness.

  • Engineering assurance

    Trace production changes from deploy back to review, tests and approval.

Roadmap items are planned, not yet available.

Give every engineer your best engineer’s context.

Book a demoTalk to us