Running UniFi Controller in High Availability Mode (HA)

You’ve got multiple sites or critical infrastructure and can’t accept controller downtime. A single controller is fine for most, but in serious setups, you need redundancy. UniFi’s High Availability (HA) mode gives you that, run two (or more) controllers in sync so if one fails, the other takes over seamlessly.

In this post, we’ll walk through how HA works, how to set it up, what to monitor, and when it makes sense versus using something like UniHosted. Let's cover it all.

Let’s dive in !!

What is UniFi controller high availability?

HA means running a cluster of UniFi OS Consoles (like UDM‑Pro, UDR, or a cloud instance) where one acts as active and the other as standby. Changes you make are replicated across both. If the active node goes offline—network issue, hardware fault, or maintenance—the standby takes control.

Clients don’t flinch. APs reconnect to whichever controller responds, and your network stays up.

Why HA matters

• No single point of failure: If your controller host crashes, clients don’t lose connectivity.
• Safe upgrades: Upgrade the standby first, test, then flip over.
• Controlled maintenance: Apply updates or patches without downtime.
• Scalable setups: Multiple geographic locations can rely on central controllers with local failover.

When HA makes sense

Consider HA if you're:

• Managing client networks or branch offices
• Running mission‑critical infrastructure (healthcare, retail, finance)
• Deploying high-density or heavy-use networks
• Unable to tolerate controller downtime

If you're running a single-site hobby setup, HA adds complexity you don’t need.

Requirements for HA

1. UniFi OS Consoles on each node (UDR, UDM‑Pro, cloud VM, or UniHosted site)
2. Same UniFi OS and Network app version
3. Controllers must see each other on LAN or secure network
4. Shared or replicated database (rocksdb) and file storage
5. Cloud access enabled for failover

UniFi OS 3.3+ has built-in support, making HA more user-friendly than past versions.

Step 1: Prepare your nodes

• Install UniFi OS Console on each machine or VM
• Assign static IPs or reserved DHCP addresses
• Enable cloud login and assign same account
• Confirm time synchronization with NTP
• Ensure firewall allows controller-to-controller communication (ports 27117, 8080, 8443, 8843, 8880, etc.)

Step 2: Join standby node to active

1. Log into your primary (active) controller’s UniFi OS dashboard.
2. Go to Settings → Controller → High Availability.
3. Click Add Node.
4. You’ll see instructions with a join token or code.
5. Log into the standby controller’s OS dashboard.
6. Paste the token to initiate sync.
7. Wait a few minutes for full database sync, devices, sites, settings.

Once complete, standby will show as Pending → Syncing → Ready.

Step 3: Validate HA status

After sync, confirm both nodes show as Healthy under HA settings.

Test failover by temporarily disconnecting the active node (power unplug or stop UniFi OS). The standby should become Active, and APs reconnect after ~30 seconds.

Re-enable the original: it comes back as Standby, syncing missed changes. UniFi keeps things smooth.

Step 4: Using HA day-to-day

With HA enabled:

• Changes to SSIDs, firewall rules, device configurations are instantly replicated
• Standby remains passive until needed, no double management
• You can schedule upgrades, reboot one node, inspect logs, all without disruption

Importantly, clients remain connected: only controller backups or migration would break them.

Step 5: Upgrade process with HA

Upgrading is safe:

1. Upgrade the standby node first.
2. Confirm everything is still Healthy
3. Manually promote standby to active (click “Promote”).
4. Upgrade the now-standby node (original).
5. Promote back or leave as-is.

You eliminate controller downtime entirely.

Step 6: Monitoring HA health

Watch these key indicators:

• HA status: Healthy, syncing, failed
• Replication time: Should be near-instant
• Logs: Look for sync errors or delays in /var/log/unifi-os/
• AP connectivity: Make sure APs show both controllers under Devices → Overview

Configure email or Slack alerts on HA health changes using OS notifications and webhooks.

Step 7: Backup and disaster recovery

HA isn’t backup. If you delete a site or misconfigure, both controllers are updated.

Plan these extras:

• Scheduled backups to cloud storage
• Export site configs periodically
• Test recoveries, can you stand up a new controller with your database?

If everything lives in UniHosted, backups and recovery are tracked for you, reducing risk.

Step 8: Common HA issues and fixes

Latency between nodes

Controllers should be under 30 ms apart in ping. Higher latency causes slow syncs or timeouts.

Fix: Use VPN or dedicated link to lower latency, or colocate controllers closer.

Split-brain errors

If both nodes think they’re Active, things break.

Fix: Reset HA, pick one to “Active”, rejoin standby cleanly.

Zabbix or other DNS redirects

If both nodes use same DNS name, APs or devices might bounce.

Fix: Use separate hostnames/IPs for management, let HA handle API traffic internally. DNS round robin is risky.

Step 9: Scaling HA clusters

UniFi currently supports 2-node clusters. The standby node can serve all sites. A third node isn’t officially supported, but UniHosted runs multiple controllers per client with global view.

For multi-site setups:

• Use a central HA cluster in cloud
• Use on-prem controllers in remote sites for uptime
• Site-to-site IPSec or WireGuard ties them together
• VPS controllers can be promoted in emergencies

Step 10: Alternatives vs. HA

If HA sounds complex, alternatives include:

• Single hosted controller (e.g., UniHosted) with built‑in failover
• Cloud Key G2 Plus with frequent backups
• Off-site cold standby, restore from backup if failure occurs

But these options involve some downtime. HA minimizes that almost to zero.

Conclusion

UniFi Controller High Availability mode brings truly resilient management to your network stack. From seamless upgrades to instant failover and mirrored settings, HA removes controller downtime from the equation. If you need that level of uptime for clients or critical business systems, setting up HA is worth the time investment.

At Unihosted, we build every hosted controller with HA in mind, and manage failover, backups, and upgrades for you. If you'd rather focus on your network and not on controller availability.