Tokonoma automatically investigates its findings and your existing alerts using telemetry, team expertise, and past investigations. Engineers and their coding agents get an explanation backed by evidence, along with proposed next steps.
From checkout failures to a proposed fix.
← →
Thread in #tk-demo-alerts
Thread# tk-demo-alerts
PagerDuty APP
Triggered: [FIRING:1] CheckoutErrorRateHigh
Alert firing on service checkout
[FIRING:1] CheckoutErrorRateHigh
Error rate above threshold in production.
Severitycritical
Servicecheckout
1 reply
Tokonoma APP
Investigated · CheckoutErrorRateHigh
Evidence points to database connections left open after checkout requests time out, exhausting the pool.
Proposed fix: Release connections when requests time out, then verify pool usage returns to normal.
Release database connections when checkout requests time out, so they return to the pool.
Investigation evidence
Connections remain in use after requests time out. Pool exhaustion lines up with the checkout failures.
Verification plan
Check connection cleanup on timeout and confirm pool usage returns to normal.
Review required before merging
01
Start with the right context
Tokonoma investigates findings and existing alerts with the context that triggered them. It checks for related work and past investigations so repeated alerts do not have to start the same investigation from scratch.
02
Understand what happened
Agents gather evidence from connected telemetry and build on earlier investigations. Findings retain their supporting context, confidence, and proposed remediations so your team can assess the conclusion.
03
Review a proposed fix
When confidence is high, Tokonoma can open a pull request with a proposed fix. Your team reviews the change before it is merged.