Agents that investigate your production systems during on-call, incidents, and daily production tasks.
Resolve AI joins every on-call rotation and starts investigating the moment an alert fires.
Agents on every on-call rotation — triaging and investigating alerts in real time.
Domain-specialized agents pursue hypotheses in parallel and work with engineers in channels.
pgdb-orders-instance-3 — migrations 0335 and 0337 were silently skipped weeks ago.event_outcome column) was merged 2026-04-24, but Drizzle's migrator skipped it due to timestamp ordering (PR #27504).column "event_outcome" does not exist and relation "order_doc_state" does not exist errors across 5 services.Background agents continuously check deployments and existing services to spot anomalies.
Watches #engineering for deploys; runs post-deploy health checks.
On deploy event
Posts daily deploy + incident digest at 7 AM PT.
Daily at 7:00 AM
Monitors orders-db-cluster for query regressions and lock contention.
On alert event
Weekly resource-utilization report across clusters and services.
Weekly · Fridays
Tells you how your code actually behaves in production directly from Slack, a terminal, or your coding agent.
What's going on with checkout-router p99 today? Is it the deploy at 10:14 PT?
Yes — drift starts ~2h after the deploy. p99 was steady at ~142ms through 12:00 PT, then climbed to ~387ms as traffic ramped.
Likely candidate: new code path in order-fulfillment.ts hits a slow query at p99 only under load. Error rate and throughput look normal — this is latency-only.
Set up an agent to watch checkout-router latency after every deploy going forward. Alert me if drift >25% within 2h.
Done. I'll watch every deploy that touches checkout-router and alert you if p99 drifts more than 25% within 2 hours.
Background agent created
Handoffs, postmortems, and runbooks written while the investigation is still fresh.
Triggered by a deploy notification in #engineering (compare URL, smoke test link, "deploy kicked off"). Post all updates in the deploy thread.
| Environment | Type | Log channel | Cluster |
|---|---|---|---|
dev0 | Pre-prod (staging) | #engineering-log-dev | dev0-cluster |
app0 | Production | #engineering-log-prod | app0-cluster |
rocket | Production (dogfood) | #engineering-log-prod | app0-cluster |
Environment progression: dev0 → app0 → rocket. Each gates the next.
{github_run_id}.{attempt}-{7_char_sha} (e.g. 24893095514.1-dac7506). The github_run_id is the dev0 build run, not the prod deploy run.Before running the health check, verify the prod deploy pipeline completed successfully for this deploy tag (Deploy to Cells workflow for app0, Deploy to Rocket from Cells for rocket).
Removing the toil of investigations, war rooms, and on-call.
“Resolve AI proved it could deliver real results in a constrained environment. It identified dependencies, surfaced accurate root causes 73% faster than our teams, all while integrating cleanly into our existing stack. ”
“Resolve AI makes our junior on-call engineers as effective as our seniors, flattening the experience curve. We’ve seen a 2x productivity lift while eliminating the runbook gap.”
“Literally everybody said Resolve AI is a ‘game changer’ when it comes to incidents and resolution”
“Resolve AI feels like a teammate who’s already done half the work. It tells me immediately if something’s critical or can wait, saving countless hours and frustration.”
“It correctly identified that all of the errors were from a single transaction being retried rather than a widespread issue. It also provides the exact method from code that the transactions failed in.”
“It was able to pinpoint the exact PR that introduced the bug and tell me the event IDs, categories, and how many listings we were over our defined rate limit.”
“We have seen compelling proof that Resolve AI arrived at the same root cause that the humans did, but 4-5 hours before an incident actually happened.”
Resolve integrates with PagerDuty and other alerting tools. When an alert fires, Resolve picks it up automatically and begins investigating.
No. Resolve joins your rotation. Engineers stay on the loop and make the final call on high-severity issues. Agents handle the investigation so engineers start with context, not a blank screen.
Engineers review every finding. When they correct the agent's reasoning, those corrections become reusable knowledge. The system gets more accurate over time.
You define the boundary. Guardrails control what agents can do on their own versus what requires engineer approval. You set the rules.
Datadog, Grafana, Prometheus, OpenTelemetry-based systems, and more. Resolve understands each tool's query language and data model.
Alert correlation groups related alerts. Resolve investigates them. It tells you why they're happening, with evidence from across your stack.
Machines on-call for humans