Skip to main content

Netshire Ltd | Technology Solutions & Digital Transformation | UK

One console for everything with superops

A problem we see time and again across IT environments, and our clients can confirm, is that issues often go unnoticed until a frustrated user picks up the phone.

This situation presented itself when we walked into a client’s Wolverhampton office: around 100 devices on site, sourced from various vendors over the years, with no way no monitor them centrally.

Servers weren’t reporting. PC’s weren’t either. The only alert was a ringing phone or Teams message.

Our goal was to fix that the right way with one dashboard and a tightly integrated ticketing system. To make that happen, we turned to SuperOps.

The approach: two routes into one console

The client’s environment required two different monitoring approaches:

  • Servers and PCs: We deployed the SuperOps agent using Group Policy.
  • Network equipment: We used SNMPv3, because network devices cannot run the SuperOps agent and must be polled instead.

Both routes ultimately delivered their information into SuperOps, giving the client one console for the whole estate.

Before changing anything, we scanned the environment and captured the existing configurations. This showed us what was actually there, including devices and software that nobody had documented. That discovery step proved valuable later, particularly when an old antivirus product became part of the deployment problem.

Deploying the SuperOps agent with Group Policy

The client runs Active Directory, so Group Policy was the most practical way to install the SuperOps agent across the domain without visiting every machine individually.

We followed SuperOps’ own GPO method. A computer startup script downloads the installer and installs it when the machine starts. The script is specific to the client and site, which determines where each device appears in SuperOps.

How we ran the deployment

We took a controlled approach:

  1. Piloted the deployment on a small group.
    We created a security group, scoped the policy to it and proved the installation on one machine before exposing it to the wider domain.
  2. Widened the scope once the pilot worked.
    After confirming the installation, we applied the policy to all domain computers.
  3. Checked permissions after changing the filtering.
    Replacing the default filtering can leave computer accounts without the required Read permission, causing the policy to fail silently everywhere. We confirmed the permissions on the Delegation tab.
  4. Verified the results from the client side.
    The script maintains its own log, and the installer also creates a log on each machine. These showed clean installation results.
  5. Re-checked the deployment later.
    This final step was important, because the script did not remain in place.

The twist: an old antivirus deleted our script

The initial pilot worked. The startup script ran and the SuperOps agent installed successfully.

Later, however, the script file had disappeared from SYSVOL. The Group Policy still listed the script, so the console looked normal, but nothing was being installed.

The cause was a legacy antivirus product that nobody had documented. A script that downloads and runs an installer is exactly the kind of activity that security software can flag.

We resolved the immediate issue by creating an exclusion for the script, allowing it to remain in SYSVOL and run as intended.

Lessons learned

This experience reinforced several important points:

  • Check for security software even if the client says there isn’t any. Old products can linger on servers, and people forget they exist. Check running services, installed software and listening ports rather than relying on memory.
  • A policy that applies is not necessarily a policy that works. Confirm that the script file exists and that the installation logs show success.
  • Check antivirus quarantine and event logs early when a script goes missing or an installation silently fails.
  • Run the script by hand to split the problem in half. If it installs manually, the script is fine and something is interfering with its delivery.
  • Name policies clearly. Two similarly named policies make it easy to edit the wrong one.
  • Consider retiring old security products. An exclusion fixes the immediate problem, but unmanaged security software is a risk in its own right.

Antivirus products often monitor startup folders and network shares for scripts that fetch and run executables. As a result, a legitimate installer script can look like a dropper.

It is also worth remembering that files in SYSVOL are replicated across domain controllers, so a deletion can spread throughout the environment.

Important deployment considerations

There are several practical details to keep in mind when deploying the agent with Group Policy:

  • A startup script only runs at boot, and installation can take 45–50 minutes afterwards, according to SuperOps. If nothing appears 20 minutes after a reboot, it may simply still be working.
  • Shutting down and powering on is not always a full boot. Windows Fast Startup can skip startup scripts, while Restart forces a real boot.
  • Laptops that only connect over VPN can be a poor fit because they may not reach the domain at boot.
  • The agent needs outbound HTTPS with TLS 1.2 to download. A proxy or firewall that only allows older TLS versions can prevent the deployment.
  • SuperOps’ guide suggests raising the Group Policy permission level to “Edit settings”. We did not do this because Read plus Apply was enough. Edit provides write access to a policy that runs code as SYSTEM at boot on every machine, so it should not be granted lightly.

Monitoring network devices with SNMPv3

For the network equipment, we used SNMPv3.

SNMPv1 and SNMPv2c send a community string in plain text. Anyone who can see the traffic can read it, and the string is often a default value or shared across multiple devices.

SNMPv3 adds proper users and can authenticate and encrypt the connection. This means credentials and monitoring data are not exposed on the wire.

A dedicated monitoring user can also be read-only and revoked independently without changing anything else.

Mixed estates require care because each vendor words its SNMPv3 settings slightly differently. Using one tested pattern consistently saves a lot of head-scratching.

How we set it up

We:

  • Created a dedicated monitoring user on the switches, with a context name set to match.
  • Added the switches to SuperOps using the SNMPv3 details.
  • Flagged the old shared v2c string for removal so it would not remain unnoticed.

Retiring Telnet and improving visibility

Telnet sends usernames and passwords in clear text, so it should be replaced with SSH or a secure web interface. Devices often ship with Telnet enabled, and it can easily be forgotten.

One lesson from this work was that a port check is not proof that a service is usable. After we disabled Telnet, a connection test still reported that port 23 was reachable. However, a real connection attempt closed without presenting a login prompt, proving it was a false positive.

Always test a service in the way a real user would.

We also set the core switch’s clock using SNTP. Correct timestamps matter because alerts and logs are difficult to trust when device clocks disagree.

The outcome

The client’s servers, PCs and network equipment are now visible in SuperOps, providing one console for the entire estate.

Insecure management protocols are being retired, and the client has a solid base for the network changes to come. Just as importantly, the discovery and monitoring work provided a clearer understanding of the environment before further changes were made.

You can’t safely change a network you can’t see, and you can’t trust what you haven’t inventoried.

If your business has a mixed-vendor network, undocumented devices or no central monitoring, our team can help you deploy the right solution and bring your environment under control.

Leave a Reply

Your email address will not be published. Required fields are marked *