Skip to content

14-day RMM trial · No card

Start free
← All guides

RMM agent recovery · 13 minute read

RMM agent offline or not checking in? Prove the failed layer before reinstalling.

Preserve the last known state, trace service and transport health, reconcile enrollment identity, and choose a supported repair or controlled reinstall without losing evidence or creating a duplicate device.

Last reviewed November 22, 2026Vendor-neutral · evidence-first workflow

First decide what “offline” actually means

The endpoint may be powered down, asleep, isolated, filtered from view, enrolled under another record, unable to reach the control plane, or running an agent whose one module is stale. Treat the red console state as a symptom. Use a durable device identity and time-aligned evidence to find the first failed layer.

The first 15 minutes

Minute 0–2

Preserve the last known state

Record UTC time, customer and site, hostname, durable device ID, console record, last check-in, last inventory, last successful job, and the operator. Confirm the endpoint is expected to be online and still managed.

Minute 2–5

Prove the service and binary

Capture the expected service state, startup type, PID, executable path, service account, exit code, agent version, SHA-256, and signer before restarting anything. Stop if the path, account, or signature is unexpected.

Minute 5–8

Collect local failure evidence

Preserve the first relevant agent error, Windows events, crash evidence, recent start attempts, timestamps, and time zone. Review excerpts for credentials and customer data before sharing them.

Minute 8–11

Test the service network path

Check DNS, the documented outbound destination and port, service-context proxy behavior, TLS negotiation, certificate chain, and endpoint time. Do not create a broad firewall or security exclusion to make the test pass.

Minute 11–15

Reconcile identity and console visibility

Match safe local enrollment identifiers to the intended tenant and site. Search for old hostnames, duplicate or replacement records, recent enrollment, inactive views, filters, and archived devices before assuming the agent is absent.

Repair or reinstall?

Prefer repair when

  • The expected package, signer, service, and enrollment identity are intact
  • A narrow service, network, certificate, clock, policy, or resource dependency is proven
  • The vendor-supported repair path preserves identity and audit continuity
  • Success can be tested after the documented ingestion interval

Consider controlled reinstall when

  • The supported repair path is unavailable or has failed with evidence
  • The package or enrollment state is irrecoverably damaged
  • Removal behavior, alternate access, authorization, and rollback are documented
  • The old and new platform records can be reconciled before automation resumes
Open the 24-control recovery checklist

Prove the full recovery path

A heartbeat is only the first acceptance check.

Nizlo offers a 14-day trial with no card and no automatic paid conversion. Use one authorized noncritical endpoint to test enrollment identity, monitoring, a bounded script, patch scan, remote support, and audit evidence before expanding a rollout.

Start the no-card trial

Continue the investigation

Offline RMM agent questions

What should I check first when an RMM agent is offline?

First confirm that the endpoint should be online and is still in managed scope. Preserve its durable device identity and last known platform state, then prove the expected service and signed binary before testing the documented service network path, enrollment identity, and console filters.

Should I repair or reinstall an offline RMM agent?

Prefer the vendor-supported repair path when the installed identity is correct and evidence identifies a narrow dependency. Use a controlled reinstall only when repair is unavailable or has failed, the state is irrecoverably damaged, and you can preserve evidence, authorize removal, control tenant placement, and reconcile records.

Does a new heartbeat mean the RMM agent is fixed?

No. Verify the correct customer and durable device identity, current inventory, expected agent version, monitoring, one bounded script, a fresh patch scan, remote support where it is in scope, and corresponding audit evidence after the normal ingestion interval.