Add and manage hosts
This guide shows you how to connect computers to LEA Platform, keep their LEA agent healthy, and find out why a host stopped reporting.
Before you start
- You need an administrator account in the console for tokens, activation, re-enrollment, the Terminal and debug capture.
- On each host you need administrator rights (root or
sudoon Linux and macOS, an elevated PowerShell on Windows). - Each host must reach the LEA Platform server over HTTPS (port 443). On an air-gapped site that is enough: the agent files are downloaded from your own LEA Platform server, at
https://<your server>/deploy/. - Words used here: a host is a computer that runs the LEA agent. A deployment token is a one-time secret that lets a new agent join. Enrollment is the moment the agent receives its own client certificate. See the glossary.
How a host joins
- You create a deployment token in the console.
- You install the LEA agent on the host with that token.
- The agent sends a certificate request together with the token. The server checks the token and gives the agent a client certificate. The agent removes the token from its configuration file.
- From now on the agent proves who it is with that certificate (mTLS). The server only talks to agents with a valid certificate.
- The host appears on the Compliance page. If the token activates hosts automatically, the host starts running the rules of its groups.
Certificates renew by themselves before they expire, as long as the agent can reach the server.
Create a deployment token
-
Open Settings → Deployment tokens.
-
Click Create Token at the top right.
-
Fill in the Create Deployment Token dialog:
Field What it does Suggestion Name A label for you, up to 64 characters. Where or how you use it, for example "Linux web servers 2026". Purpose Agent enrollment for LEA agents. Ansible / AWX gateway certificate only if you report results through Ansible or AWX. Agent enrollment. Group New hosts that use this token are added to this group. No group adds them to none. A group that has the rules these hosts need. Activate enrolled agents automatically On: new hosts start running rules right away. Off: new hosts wait as INACTIVE until you activate them. On for trusted networks. Off when you want to check every new host first. Max uses (optional) How many hosts may use the token. Empty means no limit. The number of hosts you plan to install. Expires (optional) After this date and time the token stops working. A few days after your planned rollout. -
Click Create.
You should see the dialog Deployment Token: <name> with the warning "This token is shown only once. Copy it now." It contains:
- Token: the secret itself.
- Install command for Linux and macOS (run as root on the agent host) and Install command for Windows (run in an Administrator PowerShell on the agent host): complete install commands with your server name and the token.
- Variables for bulk deployment (Ansible, GPO, Intune): the same values as separate variables.
Click the copy icon next to each item you need, store the token safely, then click Done.
Important: The console never shows the token again. If you lose it, create a new token.
Manage existing tokens
Each token is shown as a card with its state (ACTIVE, INACTIVE or REVOKED), its purpose, its group, AUTO-ACTIVATE or NEEDS APPROVAL, how many times it was used and when it expires.
- Rotate Token (circular arrows): creates a new secret. The old one keeps working for 24 hours, so running installs can finish.
- Revoke Token (bin icon): the token stops working right away. Hosts that already enrolled keep working; they no longer need the token.
Install the agent on Linux
The install script needs curl, systemd and a 64-bit Intel/AMD (x86_64) or ARM (aarch64) processor.
-
Copy the Install command for Linux and macOS from the token dialog. It looks like this:
curl -fsSL https://lea.example.com/deploy/deploy.sh | sudo LEA_SERVER='lea.example.com' LEA_ENROLL_TOKEN='<token>' bashlea.example.com: the name of your LEA Platform server, withouthttps://.<token>: the deployment token.
If your server uses its own (self-signed) certificate, the console shows a longer command with
--pinnedpubkeyandLEA_SERVER_PIN. Use that one: it makes sure the host talks to your real server. -
Log in to the host and paste the command.
You should see "LEA Deployment completed successfully on linux/amd64" (or arm64).
The script:
- creates a system user
leawith no login shell, - installs the agent in
/var/lib/lea(the agentlea-agent-linux-<arch>, the runnerlea-agent-runner-linux-<arch>and the configuration filelea.conf), - writes your server address and the token into
lea.conf, - creates and starts the systemd service
lea.
Note: The script stops with "/var/lib/lea Exists" if the agent was installed before. To reinstall, remove the old agent first: run
curl -fsSL https://<your server>/deploy/undeploy.sh | bash. It prints the commands that stop the service and remove the user and the folder; check them, then run them.
Note: The script first gets the server's signing key over a verified connection. If that fails, it stops before it changes anything on the host. Check that the host trusts the server's certificate, or use the console's command with
LEA_SERVER_PIN.
Note: The
leauser gets nosudorights. If a rule needs one privileged command, its description names the exact line to add to a file in/etc/sudoers.d/.
Check the agent:
sudo systemctl status lea
sudo journalctl -u lea -n 50
Hosts with SELinux enforcing (RHEL, Rocky, AlmaLinux, Oracle Linux, Amazon Linux)
The script installs the agent in /var/lib/lea, and files there get the SELinux type var_lib_t. With SELinux in
Enforcing mode, systemd is not allowed to start a program of that type: the service lea restarts again and again,
and journalctl -u lea shows Failed to locate executable /var/lib/lea/lea-agent-linux-amd64: Permission denied and
status=203/EXEC. The host never appears in the console.
Check the mode with getenforce. If it prints Enforcing, run this once on the host after the install script (as root):
semanage fcontext -a -t bin_t '/var/lib/lea/lea-agent(-runner)?-linux-[^/]+'
restorecon -v /var/lib/lea/lea-agent*-linux-*
systemctl restart lea
semanage fcontextrecords the rule, so a relabel of the file system keeps it;restoreconapplies it to the files that are there now.semanagecomes withpolicycoreutils-python-utils(RHEL 8 and later).- Check:
ls -Z /var/lib/lea/lea-agent-linux-*showsbin_t, andsystemctl status leashowsactive (running). - Do not switch SELinux off (
setenforce 0) for the agent. - After an agent update (see Update the agent) the new program file gets
var_lib_tagain: run therestoreconandsystemctl restart lealines again if the service does not stay running.
Install on many Linux hosts
Use the Variables for bulk deployment from the token dialog in your deployment tool (for example an Ansible task) and run the same deploy.sh command with them: LEA_SERVER, LEA_ENROLL_TOKEN and, with a self-signed server certificate, LEA_SERVER_PIN.
Install the agent on macOS
The same deploy.sh script installs the agent on macOS. Only Apple silicon (arm64) is provided.
- Copy the Install command for Linux and macOS from the token dialog.
- Open Terminal on the Mac and paste the command. Enter your administrator password when
sudoasks.
You should see "LEA Deployment completed successfully on darwin/arm64".
The script:
- creates a hidden system user
leawith no login shell, - installs the agent, the runner and
lea.confin/var/lib/lea, - writes your server address and the token into
lea.conf, - creates and starts the launchd service
com.lea.agent(/Library/LaunchDaemons/com.lea.agent.plist).
The agent writes its messages to /var/lib/lea/lea-agent.log. To remove it, run curl -fsSL https://<your server>/deploy/undeploy.sh | bash, check the printed commands, then run them.
Install the agent on Windows
The PowerShell script deploy.ps1 installs the agent on 64-bit Intel/AMD (amd64) and ARM (arm64) Windows. It needs curl.exe, which ships with Windows 10 and later.
-
Open PowerShell as Administrator.
-
Copy the Install command for Windows from the token dialog and paste it. It looks like this:
$env:LEA_SERVER='lea.example.com'; $env:LEA_ENROLL_TOKEN='<token>'; curl.exe -fsSL https://lea.example.com/deploy/deploy.ps1 | Out-String | Invoke-ExpressionWith a self-signed server certificate, the console's command also sets
LEA_SERVER_PINand uses--pinnedpubkey.
You should see "LEA Deployment completed successfully on windows/amd64" (or arm64).
The script:
- installs the agent, the runner and
lea.confinC:\ProgramData\LEA(set$env:LEA_INSTALL_DIRfirst to use another folder), - writes your server address and the token into
lea.confand limits the folder to SYSTEM, Administrators and the service account, - installs and starts the service LEA Agent (
LEAAgent). It runs as the service accountNT SERVICE\LEAAgent, which is not an administrator.
Note: The service needs the Log on as a service right. If a group policy sets that right, make sure the policy still includes
NT SERVICE\ALL SERVICES.
Note: Commands that need administrator rights fail under this account. In the rule library this affects Disk Encryption Enabled (Windows), because
manage-bdeneeds an administrator.
The agent writes its messages to lea-agent.log in the same folder; follow it with Get-Content -Wait C:\ProgramData\LEA\lea-agent.log. The script stops with "Exists" if the folder is already there. To remove the agent, run curl.exe -fsSL https://<your server>/deploy/undeploy.ps1 | Out-String | Invoke-Expression, check the printed commands, then run them.
For GPO or Intune, set LEA_SERVER, LEA_ENROLL_TOKEN and, if needed, LEA_SERVER_PIN from Variables for bulk deployment and run deploy.ps1 as SYSTEM.
Remove the old rights from existing hosts
Agents installed by an older version keep their old rights until you update them with the install script. An automatic agent update does not change them, because it only replaces the agent program.
Run the command from Settings → Agent updates → Update an agent by hand on each host:
- Linux: removes the
sudoline the older installer added to/etc/sudoers.d/lea. Lines you added yourself stay, and the file is deleted if nothing else is left in it. - Windows: moves the service from SYSTEM to
NT SERVICE\LEAAgentand gives that account access to the agent folder.
Turn on server checks on existing hosts
Agents installed by an older version can have LEA_SKIP_SSL_VERIFY=true in lea.conf, and some have no LEA_SERVER_SIGNING_KEY. The same update by hand fixes both: it turns the certificate check on, adds your pin if the command has one, and writes the server's signing key. If the host cannot verify the address in its lea.conf, the update stops before it changes anything; add LEA_SERVER_PIN to the command.
Without a signing key, a new agent version does not start. An automatic update then ends in FAILED and the host stays on its old version until you run the update by hand.
Agent configuration file
The agent reads lea.conf from its own folder (Linux: /var/lib/lea/lea.conf). Restart the agent after you change it (Linux: sudo systemctl restart lea; Windows: Restart-Service LEAAgent).
| Setting | Meaning | Default |
|---|---|---|
LEA_SERVER_URL | Address of the server, ending in /lea. | set by the installer |
LEA_ENROLL_TOKEN | Deployment token. Removed by the agent after it enrolls. | empty |
LEA_SKIP_SSL_VERIFY | true turns off the check of the server certificate. Use only for tests. | false |
LEA_SERVER_SIGNING_KEY | Public key the agent uses to check every server reply. The agent does not start without it. | set by the installer |
LEA_SERVER_PUBKEY_PIN | Pin of the server certificate, for self-signed certificates. | empty |
LEA_INTERVAL | Seconds between check-ins with the server. | 30 |
LEA_COMMAND_TIMEOUT | Maximum seconds a command may run. | 60 |
LEA_DEBUG | 1 writes detailed debug messages to the agent's own log. | 0 |
Activate or deactivate a host
An inactive host is not evaluated and not counted in compliance numbers. Hosts that enroll with a token without Activate enrolled agents automatically start inactive. They show the pill INACTIVE on the Compliance page and on the host page.
You need permission to manage hosts. Use one of these ways:
-
Host page: open the host and click Activate at the top right. On an active host the button is Deactivate; confirm it in the dialog.
-
Several hosts: on the Compliance page, click the select icon in the first column header, tick the hosts and click Activate or Deactivate under the table.
-
AI assistant (when an administrator turned it on): ask "Activate host web-01" (or "Deactivate host web-01"). The assistant proposes the change; click to approve it.
-
LEA command line: find the host's Client ID on its host page (Overview tab), then run:
lea api -X PATCH "clients?id=eq.21" -d '{"active": true}'Replace
21with the Client ID. Usefalseto deactivate. To activate several hosts at once, list them:"clients?id=in.(21,22,23)".
The change is recorded in the audit log under your name. If your organization requires approval for host actions, the change waits for an approver first. You should see the INACTIVE pill disappear when the host is active.
To avoid manual activation, turn on Activate enrolled agents automatically on the token.
Fix enrollment problems
The Enrollment column on the Compliance page and the Enrollment panel on the host's Overview tab show the identity state of each agent. The states are listed in Check compliance.
Agents that fix themselves
If an agent still has its private key but its certificate expired, was lost or is no longer known to the server (for example after a database restore), it asks for a new certificate and proves it holds the old key. The server issues the new certificate automatically. You do nothing.
Re-enroll a host that lost its identity
Use this when the agent lost its key (files deleted, disk replaced, cloned virtual machine) and has no deployment token, or to move an upgraded LEGACY host to its own certificate. Nothing needs to be done on the host.
-
Open the host page, Overview tab, and click Re-enroll in the Enrollment panel.
For many hosts: on the Compliance page, filter Enrollment → Certificate error, click the checkbox icon (Select mode) in the table header, select the hosts and click Re-enroll under the table.
-
In the Re-enrollment dialog, choose Selected hosts or All hosts matching the current filters.
-
Set Window (minutes): how long the host may get a new certificate without a token (1 to 1440 minutes; default 60).
-
Click Open Window.
You should see "Re-enrollment window opened for N host(s)." The Enrollment state changes to RE-ENROLLMENT OPEN.
What happens next:
- The old certificate stops working immediately.
- At its next check-in (usually under a minute) the agent asks for a new certificate and gets it. Only one request is accepted per window.
- The host moves to PROBATION. It runs rules again, but rules that use secret parameters (passwords, keys) are held back. Its history and groups are kept.
- Check that it is the right machine: compare the IP address and system details on the host page with what you expect.
- Open the Re-enroll dialog again for the host and click Confirm.
You should see "N host(s) confirmed and released from probation." The state returns to mTLS enrolled.
Close Window cancels a window that was not used yet.
Warning: RE-ENROLLMENT CONFLICT means a second machine asked for a certificate for the same host during probation. One of them may not be your real host. Find out which machine holds the certificate (IP address on the host page, and the audit log) before you click Confirm. If it is the wrong one, open a new window and make sure the real agent is running.
Move old agents to certificates
Old agents (shown as LEGACY) do not use a certificate. The server only accepts them while Require agent certificates is off.
-
Open Settings → Security. Under Agent certificates, turn off Require agent certificates and confirm it is you.
You should see a red box: "Agent certificates are not required." Old agents now check in and show as LEGACY.
-
Update the agents to the current version: see Update the agent.
-
Re-enroll the updated hosts: see Re-enroll a host that lost its identity. On the Compliance page you can filter Enrollment → Legacy and re-enroll all of them at once.
-
Check and Confirm the hosts in PROBATION.
-
When no host shows LEGACY any more, turn Require agent certificates back on. The red box disappears.
Warning: While the setting is off, new agents can register without a deployment token. Keep this period short.
Renew the agent certificate authority
All agent certificates are signed by the server's agent certificate authority (CA). When it expires, every agent that uses its certificate is rejected.
-
Open Settings → Security and look at Certificate authority.
You see the Agent CA and the Server HTTPS certificate with their expiry date. EXPIRING (amber) means less than 90 days are left, EXPIRED (red) means it already expired. The Security card on the Settings overview turns the same color.
-
Click Renew agent CA and confirm it is you.
You should see the Agent CA with a new expiry date 10 years from now. Agents keep their certificates; nothing changes on the hosts. In a cluster every node takes the renewed CA within a minute.
The server's HTTPS certificate is not renewed here. Replace it, or let the cluster renew it automatically (see High availability).
Hosts added through the Ansible/AWX gateway
A host that registers itself through the Ansible/AWX gateway starts in PROBATION: it runs its rules, but rules that use secret parameters are held back. Check that it is the right machine, then open the Re-enroll dialog for the host and click Confirm.
More detail: Agent enrollment reference.
Use the host page
Open a host by clicking it anywhere in the console and then Open host page, or Ctrl/⌘-click its name. The header shows the compliance status, the worst failing severity, INACTIVE when switched off, and a line with the operating system, IP address, agent version and when it was last seen. The time turns orange when the host has not been seen for over a day.
Overview
- Compliance posture: the same summary as the host preview, with failing, awaiting and passing rules.
- Host: Client ID (the number used in addresses and the API), LEA GUID (the agent's identity), Registered, Last seen, Check-in interval.
- System: operating system, distribution, kernel, architecture, IP addresses and Agent version.
- Enrollment: state, certificate state, Deployment token, Certificate expires, the last rejection with its reason, and the Re-enroll button.
- Collected host tags: values that rules collect about the host, for example its host name. Click a row to see the full output.
| Column | What it shows |
|---|---|
| Tag | The name under which a rule stores the value it collects, set as the rule's Command info key (tag). JSONPath groups refer to it, for example $.hostname.stdout. |
| Value | The latest output of the rule that collects this tag. |
| RC | The return code (exit code) of the collecting command in its latest run; 0 usually means the command ran without error. |
| Stderr | Error messages the collecting command printed in its latest run (standard error). Empty when there were none. |
Rule results
Every rule the host runs with its latest status. Use the buttons All, Failing, Passing and Info, or Filter rules to search. Click a rule to see its result. See Read results and fix failures.
| Column | What it shows |
|---|---|
| Status | Whether the latest run of the rule on this host met the rule's expectations: PASS, FAIL or INFO. |
| Rule | The rule. The version chip on the right shows the version the latest run used, only for rules that have more than one version; it is orange when the rule has a newer version now. |
| Severity | How serious it is when this rule fails, as set in the rule, from Critical down to Informative. |
| RC | The return code (exit code) of the rule's command in its latest run; 0 usually means the command ran without error. |
| Output | What the rule's command printed in its latest run, or its error messages when it printed nothing else, with all lines joined into one line. |
| Last check | When the rule last ran on this host. "never" means it has not run here yet. |
| Assignment | Whether the host still runs this rule through one of its active groups. Rows that are "not assigned" are greyed out and show an old result. |
Run history
Every run of every rule on this host. Turn on Status changes only to see only the moments a verdict changed.
Agent logs
Messages the LEA agent sent to the server.
- Filter by level: All, Errors, Warnings, Info, Debug. Use Search messages to search.
- Click the follow button to see new messages as they arrive. Switch between Table and Raw.
| Column | What it shows |
|---|---|
| Level | How important the agent rated the message, from DEBUG (detail for troubleshooting) through INFO and WARNING up to ERROR and CRITICAL. |
| Message | The text the agent wrote. Click the row to see all of it. |
| Agent time | When the agent wrote the message, by the host's own clock. |
| Received | When the server received the message from the agent, usually at the agent's next check-in. |
To capture debug messages (administrators):
- Click Capture debug logs.
- Choose For 5 minutes, For 15 minutes, For 30 minutes or For 60 minutes.
You should see the pill CAPTURING DEBUG with the time left (see the table below). Capture starts at the agent's next check-in, and the host checks in faster meanwhile. You get a "Debug logs arriving" notification when the first messages arrive. Click +15 min to extend, or the x (Stop capturing debug logs) to stop. Captured messages appear under the Debug filter.
| Pill | Meaning |
|---|---|
| CAPTURING DEBUG | The agent sends its detailed DEBUG messages to the server until the time shown, and checks in faster meanwhile. |
Agent logs are kept for 90 days and debug messages for 7 days by default. See Settings → Data retention.
Terminal
Run a one-off read-only command on the host and see its output. See Run a one-off command on a host.
| Pill | Meaning |
|---|---|
| FAST CHECK-INS | The agent contacts the server more often than usual until the time shown, so commands and debug logs arrive sooner. |
Groups
The groups this host belongs to and how, with the number of rules each group has. Administrators can:
- add the host to a group with + Add to group... at the bottom of the table,
- remove it with the remove icon. This only works for groups the host was added to directly; for other memberships, change the group that provides it.
| Column | What it shows |
|---|---|
| Membership | How this host belongs to the group: added by hand ("direct member"), through a nested group, by a JSONPath expression ("dynamic") or as "all hosts". |
| Rules | How many rules were added to the group itself, not counting rules it takes from nested groups. |
See Groups.
Held rules
Rules this host runs at an older version, because a rule change was rolled out to new hosts only or because a group is pinned to a version. Click Update to move the host to the current version, or Update all. See Rule library and updates.
On-host self-service (force a rule)
An administrator on the host can run one rule right away, by its name, without signing in to LEA Platform. On Linux:
cd /var/lib/lea && sudo -u lea ./lea-agent-runner-linux-amd64 --force-cmd='<rule name>'
Use arm64 instead of amd64 on ARM hosts. On Windows, run .\lea-agent-runner-windows-amd64.exe --force-cmd=<rule name> in an elevated PowerShell in the agent folder.
This is intended: a host owner can, for example, join the host to an Active Directory domain with a rule whose password is a secret parameter, without seeing that password in LEA Platform.
- The agent checks in once, runs only that rule and exits. Any active rule with that name works, also when it is not assigned to the host's groups. A host in PROBATION does not get rules with secret parameters.
- Secret parameters are filled in before the command is sent to the host, so the secret value reaches the process on the host. Anyone who is an administrator on the host could capture it there.
- The result appears on the host page like any other result.
LEA Platform never shows secret values in results or agent logs: every exact copy of a secret value (4 characters or longer) in the output or in an agent log message is replaced by ******** before it is stored. A value that the command changes (for example base64-encoded) is not recognised. More detail: Agent enrollment reference.
Update the agent
Agent versions are managed in Settings → Agent updates: import a version, pin it to a test group, then make it the fleet default.
- On its next check-in, a host whose assigned version differs downloads the program for its system from the LEA server and checks its SHA-256 checksum.
- Agents from 2.1.0 restart themselves into the new version. If it does not start, they go back to the old version.
- Agents older than 2.1.0 download the new version and switch at their next restart. Linux:
sudo systemctl restart lea. Windows:Restart-Service LEAAgent.
On hosts with SELinux enforcing, label the new program file after the update: Hosts with SELinux enforcing.
Check the result on the host page: Overview → Agent version, or the agent version in the line under the host name.
Full steps, rollback and updating by hand: Agent updates.
A host stops reporting
LEA Platform sorts active hosts by how recently they reported. The dashboard widget Check-in freshness counts them:
| Label | Meaning |
|---|---|
| Checking in | Reported within about three check-in intervals (at least 5 minutes). |
| Late (< 24h) | Missed some check-ins, but reported within the last day. |
| Stale (> 24h) | No report for more than a day. |
| No results yet | Never sent a result. |
The host page → Terminal tab shows a similar status for one host, based on its last check-in: REACHABLE, LATE, STALE or NEVER SEEN. A stale host usually means the agent stopped, the host is switched off, the network or a proxy blocks it, or the server rejects its certificate.
Work through this list from top to bottom.
-
When was it last seen? On the Compliance page, sort by Last seen. On the host page the "seen ... ago" text turns orange after a day.
-
Is the identity rejected? Look at the Enrollment column and the host's Overview → Enrollment panel. A red box "Last rejected <time>: <reason>" means the server refuses the agent. Re-enroll it: see Re-enroll a host that lost its identity.
-
Does the agent run?
- Linux:
sudo systemctl status leashould sayactive (running). If not:sudo systemctl restart lea. - Windows:
Get-Service LEAAgentshould sayRunning. If not:Start-Service LEAAgent. - macOS:
sudo launchctl list | grep com.lea.agent.
- Linux:
-
What does the agent say? Read its log.
- Linux:
sudo journalctl -u lea -n 100 - Windows:
Get-Content -Tail 100 lea-agent.login the agent folder. - macOS:
tail -n 100 /opt/lea/lea-agent.log
When the runner stops, it exits with a code: 1 means it was started as root, 2 a configuration error, 3 a certificate error, 4 a network error. "no server signing key configured" means
lea.confhas noLEA_SERVER_SIGNING_KEY: see Turn on server checks on existing hosts. - Linux:
-
Can the host reach the server? On the host run
curl -sS https://<your server>/lea/server_identity. You should see a short JSON text. Errors point to DNS, a firewall or a proxy between the host and the server. With a self-signed server certificate, add-kfor this test only. -
Is the clock right? Certificates need the correct time. Compare
dateon the host with the server and fix time synchronisation. -
Was the host renamed or cloned? The agent identifies a machine by an operating system ID plus its host name. After a rename the host appears as a new host and needs a deployment token in
lea.confto enroll; the old entry stops reporting. -
Did the agent lose its files? The agent keeps its key and certificate as hidden files ending in
.keyand.crtin its folder (Linux:/var/lib/lea). If they are gone, re-enroll the host. -
Is the host inactive or in no group? Then it reports, but runs no rules. See Activate or deactivate a host and Groups.
If something goes wrong
- The install script fails with "Unsupported architecture": the host is not 64-bit Intel/AMD or ARM.
- The install script fails to download: the host cannot reach
https://<your server>/deploy/. Check the name, DNS and firewall. With a self-signed certificate, use the pinned command from the token dialog. - The host does not appear after install: the token may be expired, revoked or used up (check the token card), or the agent cannot reach the server. Read the agent log.
- The host appears as INACTIVE: the token does not activate automatically. Activate the host.
- The host shows NEVER CONNECTED: it received a certificate but never used it. Check that the agent runs and reaches the server.
- "Refusing to run as root" in the agent log: start the agent as a normal user, or on macOS add
--user=<account>as shown above.
For more, see Troubleshooting.
Related
- Check compliance
- Groups
- Read results and fix failures
- Agent enrollment reference
- Agent updates
- Agent auto-update reference
- Agent version reference
- Installation and upgrade