01
Monitoring alert
An NMS detects a symptom
Rez receives the alert, identifies the affected site and service dependencies, and begins scoped triage instead of forwarding another notification.
Rez Network Advisor
Dashboards don't fix networks. Engineers do.
Rez is the always-ready, read-only network advisor on your team. When an alert arrives, a Netcode verification fails, or an engineer asks a question, Rez scopes the affected service, gathers fresh SSH and API evidence, and explains the root cause with proof.
Rez investigates. Engineers decide. When the next safe action is confirmed, Rez prepares a governed Netcode recommendation for review. Rez never applies configuration.
Fresh live evidence
Read-only device access
Human-approved remediation
Rez Operations · Incident INC-2048
Read-only
SWSolarWinds alert · application path unavailable
Users at Branch 204 cannot reach the payments service. Find the root cause and the next safe action.
- 01Service and dependency scope resolvedBranch 204 → WAN → Edge-FW-01 → paymentscomplete
- 02Fresh device evidence collectedRouting, policy, NAT, interfaces, and return pathlive
- 03Competing hypotheses testedForward policy passes; expected return route is absentproven
Root cause confirmed
Return route missing at the WAN edge
The firewall accepted the forward flow, but the branch prefix was not present on the return path. The service failed after the controller reported success.
One advisor, three ways to start
Alerts arrive. Automation fails. Engineers ask. Rez starts the same investigation.
Every entry point resolves to the same governed workflow: understand the scope, collect fresh evidence, validate the cause, and recommend the next safe action.
02
Failed verification
A Netcode check exposes an exception
Rez inherits the change, target, expected result, and verification evidence so the investigation starts with operational context already attached.
03
Engineer question
An engineer asks from the tools they use
Start from Ops Dashboard or Slack with a plain network question. Rez narrows the scope and gathers the read-only facts needed to answer it.
The agent investigates. The math proves it.
A conclusion your network team can inspect, challenge, and verify.
Rez does not stop at a plausible explanation. It links each conclusion to fresh device evidence, checks the dependency chain, and separates the root cause from downstream symptoms.
Resolve the real scope
Map the affected site, devices, protocol adjacencies, service path, and approved expectations before collecting evidence.
Read live state locally
The Local Connector runs bounded, read-only SSH and API checks with credentials that stay inside the customer environment.
Test competing causes
Routing, L2, firewall policy, NAT, interfaces, and return path are reconciled instead of treating the first alert as the answer.
Publish the evidence chain
The final RCA shows what failed, why it failed, the affected dependency, and the evidence behind the recommendation.
Branch
POS
POS
WAN
edge
edge
Firewall
policy
policy
Payments
service
service
Forward routeBranch prefix resolves through the approved WAN pathpass
Firewall policyExact TCP/443 flow matches the installed allow rulepass
NATExpected source translation is presentpass
Return pathBranch prefix is absent from the upstream routing tableblocked
Confirmed root cause
Application return path is missing after a successful policy install.
Each finding remains linked to its device, command or API result, timestamp, and investigation record.
From alert to answer
Your monitoring platform tells you what changed. Rez tells you why service failed.
Keep SolarWinds, ThousandEyes, and the monitoring systems your team already uses. Rez turns their alerts into an active investigation that crosses routing, SD-WAN, firewall, NAT, and return-path boundaries.
SolarWinds
ThousandEyes
Monitoring webhooks
Netcode verification
BGP neighbor down at Branch 204
The alert identifies the symptom and device. Dependent service checks are also failing.
Rez investigates live
Peer shutdown in configuration, not a carrier failure
The configured state, operating state, affected routes, and downstream reachability are reconciled in one evidence-backed record.
Work where network engineers work
Ask once. Follow the same investigation from alert to recommendation.
Rez keeps the scope, live evidence, findings, and recommendation attached to one incident. Engineers can engage from the operating surface that fits the moment without restarting the investigation.
Ops Dashboard
Ask a network question, follow live reads, inspect the evidence, and review the final RCA in one workspace.
AvailableSlack
Start from an operations thread and return the evidence-backed answer to the team handling the incident.
AvailableMicrosoft Teams
Bring the same investigation workflow into Teams without creating a separate troubleshooting process.
PlannedFollow the failure end to end
One service can fail across six different network boundaries.
Rez evaluates the layers that determine whether the application works, rather than declaring success because one controller completed its job.
Interfaces and adjacency
Link state, LLDP, trunks, port channels, and the first broken dependency.
Forward path
Routes, next hops, protocol state, and propagation across the scoped path.
ACL and firewall
The exact flow verdict and the installed policy that governs it.
Translation
Expected address translation and the state required for the return flow.
Reverse path
The destination-to-source route and asymmetric-path conditions.
Packet evidence
Packet-trace facts that distinguish network loss from endpoint backpressure.
Close the work loop
When Rez confirms the next safe action, Netcode takes over.
Rez does not become an autonomous configuration engine. A confirmed recommendation becomes a governed Netcode draft with exact scope, commands, rollback, dry-run, and human review.
01
Rez confirms the dependencyRoot cause, evidence, target, and recommendation remain linked.
read-only02
Netcode prepares the changeExact commands, rollback, dry-run, target scope, and audit record.
draft03
The engineer decidesReview, approve, apply, verify, or reject. No write occurs without the human.
approvalRez Network Advisor
Give every alert and automation exception an engineer-grade first response.
Start with read-only investigation. See the evidence. Decide whether the recommendation should move into Netcode.