---
title: "Welcome to Guardian Essentials"
description: "Get started with Guardian Essentials: add your CIs, scan them, review drift and check compliance, with what to expect at each step."
product: "Guardian"
edition: "Essentials"
type: "get-started"
url: "https://docs.cloudhouse.com/guardian/essentials/get-started/"
---
# Welcome to Guardian Essentials

A practical first run through Guardian Essentials. By the end you'll have your own CIs connected, scanned, reporting changes and checked against a compliance benchmark.

Applies to: `Guardian Essentials`

!!! tip "In a hurry?"

    The short version is four steps: [add a CI](#s4), [scan it](#s6), [review what changed](#s8) and [approve what you expected](#s10). The rest of this page explains each one, and what to look out for.

## How to use this guide

This guide takes you from a new Guardian Essentials organisation to monitoring part of your own estate. Work through it in order the first time. Each section starts with a question you're likely to have and ends with something you do in Guardian. The detailed steps for each screen are on the pages it links to.

The screenshots use example data. Your CI names, numbers and settings will be different.

### Before you start

- Access to your Guardian Essentials organisation. If you've been invited, use the link in your invitation email to set a password, then sign in.
- At least one device you're allowed to scan. A test or non-production server is best for a first run.
- Either permission to install software on that device (for the agent), or an account on it that can read its configuration (for agentless scanning).
- For agentless scanning only: a machine inside your network where you can install a Connection Manager, with a network route to the devices you want to scan.

### How long it takes

| Sections | What you'll have at the end | Time |
| --- | --- | --- |
| [1–2](#s1). What this is for | The terms you need, and what Guardian is for | 10 minutes reading |
| [3–6](#s3). Getting your estate in | One CI scanning, then the rest added in bulk | 1 to 2 hours |
| [7–8](#s6). First results | A scan and your first detected changes | About an hour, plus time between scans |
| [9–11](#s9). Day-to-day questions | Comparisons, change review and a compliance report | About an hour |
| [12](#s12). Next steps | How Guardian grows with you | 10 minutes reading |

Sections 7 and 8 need a little patience. Guardian shows changes by comparing one scan with the next, so you need at least two scans before there's anything to see. [Section 8](#s8) shows you how to make a safe test change so you don't have to wait.

## 1. The problem Guardian solves

Most teams find out about an unplanned change when something breaks. A setting on a server was changed weeks or months ago. Nobody recorded it, and the first sign was an application misbehaving. Most of the cost comes after that, working out what changed, on which server, and when.

The same problem usually shows up in one of three ways.

**You can't tell what changed.** A server isn't behaving the way it did last month, and there's no record of how it was configured last month.

**Servers that should match don't.** Two web servers were built from the same image. One of them now fails intermittently, and finding the difference by hand means reading two configurations line by line.

**You can't show an auditor the evidence.** You're asked to show that your servers meet a recognised hardening standard. You think they do, but collecting the evidence by hand takes days.

Guardian deals with all three in the same way. It scans your systems on a schedule, records how they're actually configured, and tells you what changed between scans. It also checks each system against policies and benchmarks and gives it a compliance score.

## 2. Important terms in Guardian

Guardian uses the same words as your CMDB: a **CI** is anything Guardian scans, a **scan** records its configuration, and **drift** is what changed between two scans. [Guardian concepts](https://docs.cloudhouse.com/guardian/essentials/get-started/concepts/index.md) explains every term the product and these pages use, in one place. Read it once now. The terms will make more sense when you see them in use in sections 7 and 8.

!!! note "Two different questions"

    Drift answers "what has changed on this CI?" Policies and benchmarks answer "does this CI meet the standard?", whatever has or hasn't changed. You'll use both.

## 3. Finding your way around

Everything in Guardian is reached from the menu on the left. It's split into two groups: **Monitoring**, for what Guardian has found, and **Configuration**, for how it scans.

| Menu item | What it's for | More |
| --- | --- | --- |
| Dashboard | A summary of your estate, and what needs your attention | [Dashboard](https://docs.cloudhouse.com/guardian/essentials/dashboard/index.md) |
| All CIs | Every CI, with its score, issues, drift and status. Where you add and compare CIs | [CIs](https://docs.cloudhouse.com/guardian/essentials/cis/index.md) |
| Scan Results | Check results across all your CIs, grouped by severity | [Scans](https://docs.cloudhouse.com/guardian/essentials/scans/index.md) |
| Detected Drift | Every change Guardian has found, and where you review and approve it | [Detected Drift](https://docs.cloudhouse.com/guardian/essentials/detected-drift/index.md) |
| Benchmarks | The CIS benchmark library, and which benchmarks are assigned | [Benchmarks](https://docs.cloudhouse.com/guardian/essentials/compliance/benchmarks/index.md) |
| Policies | The policies your CIs are assessed against | [Policies](https://docs.cloudhouse.com/guardian/essentials/compliance/policies/index.md) |
| Reports | Compliance reports you can generate and export | [Reports](https://docs.cloudhouse.com/guardian/essentials/compliance/reports/index.md) |
| Scan Configuration | When scans run, set per CI group or environment | [Scan schedules](https://docs.cloudhouse.com/guardian/essentials/scans/schedules/index.md) |
| CM Groups | Connection Manager groups, used for agentless scanning | [Connection Managers](https://docs.cloudhouse.com/guardian/essentials/scans/connection-managers/index.md) |
| Credentials | Saved accounts that Connection Managers use to scan CIs | [Credentials](https://docs.cloudhouse.com/guardian/essentials/scans/credentials/index.md) |
| Integrations | Connections to other services, such as ServiceNow and AWS | [Integrations](https://docs.cloudhouse.com/guardian/essentials/integrations/index.md) |

Along the top are a search box for CIs, scans and changes (press `⌘K` or `Ctrl+K`), a light and dark theme switch, and your account menu.

![The Dashboard](https://docs.cloudhouse.com/guardian/essentials/assets/screenshots/dashboard.png)

Most pages are empty until you've added and scanned a CI, so don't worry if the Dashboard shows zeroes. Until you add your first CI, it shows **Welcome to Guardian**, with the quickest ways to add some.

## 4. Adding your first CI

Get one CI scanning before you add any more. Once you've seen a working CI, a broken one is much easier to spot.

Go to **All CIs** and select **Add CIs**, then **Add Manually**. The bulk options are covered in [section 5](#s5).

![The Add CI dialog](https://docs.cloudhouse.com/guardian/essentials/assets/screenshots/add-a-ci.png)

### Agent or agentless

The main choice is how Guardian should scan the CI.

| | Install the Guardian Agent | Agentless scanning |
| --- | --- | --- |
| How it works | A small agent on the CI reports directly to Guardian | A Connection Manager on your network connects to the CI over WinRM (Windows) or SSH (Linux and macOS) |
| On the CI | The agent | Nothing |
| Network | The agent connects out to Guardian. Nothing needs to connect in to the CI | The Connection Manager needs a route to the CI on the WinRM or SSH port |
| What you need | Permission to install software on the CI | A [Connection Manager](https://docs.cloudhouse.com/guardian/essentials/scans/connection-managers/index.md), and an account on the CI with read access |

The agent is usually the quicker route for a first CI. Use agentless when you can't install software on the CI. [Add CIs](https://docs.cloudhouse.com/guardian/essentials/cis/add-cis/index.md) has the steps for both.

- **With the agent,** the install steps include the details the agent needs to register with your organisation, so treat them like a password and don't share them. The CI appears on **All CIs** once the agent registers.
- **Without an agent,** use a service account rather than a person's own login. If that person leaves or changes their password, scans stop and the cause won't be obvious. Save the account under [Credentials](https://docs.cloudhouse.com/guardian/essentials/scans/credentials/index.md) if you'll use it for other CIs.

### Check it's working

Go back to **All CIs**. Each CI shows whether it's online and when it was last scanned. If a scan failed, the reason is shown at the top of the CI's page.

![All CIs](https://docs.cloudhouse.com/guardian/essentials/assets/screenshots/all-cis.png)

!!! note "If an agentless CI won't scan, check these in order"

    1. **The Connection Manager.** Open **CM Groups** and check the group shows its Connection Manager as connected. A group can be active while its Connection Manager is offline.
    2. **Network route.** The Connection Manager can't reach the CI because of a firewall, routing or the wrong network segment. Test from the Connection Manager's machine, not from your own laptop.
    3. **Credentials rejected.** The username or password is wrong, or the account isn't allowed to sign in remotely.
    4. **Credentials accepted, but without enough access.** The scan runs but some sections come back empty, because the account couldn't read them. If a scan looks thinner than expected, check permissions first.

## 5. Adding CIs in bulk

Adding CIs one at a time is fine while you're learning, but not for a whole estate. **Add CIs** also discovers CIs in your cloud accounts and imports them from your CMDB, for example from AWS or your ServiceNow CMDB. [Add CIs](https://docs.cloudhouse.com/guardian/essentials/cis/add-cis/index.md) lists every option available to you and when to use each.

Whichever you use:

- **Start small.** Import a few CIs and check they arrived as expected before you bring in the rest. A mistake is quick to fix on five CIs and slow to fix on five hundred.
- **An import is only as good as its source.** Guardian imports what your CMDB or asset system says you have, so if it's out of date, the import will be too.
- **Count what arrived.** Compare it with what you expected. An import that only partly worked looks much the same as one that fully worked.

## 6. Reviewing detected CIs

**All CIs** has two tabs: **Managed**, the CIs Guardian scans, and **Detected**, CIs that were discovered or imported but aren't managed yet.

!!! note "Detected CIs aren't scanned until you manage them"

    They don't count towards your compliance score or drift, and they stay in the **Detected** tab until you act on them.

This catches people out. The import appears to work, but nothing gets scanned, because the CIs are waiting for review. Open the **Detected** tab and decide, for each one, whether to start managing it or leave it out on purpose. Leaving a CI out on purpose makes it clear it was a decision, not something that was missed.

## 7. Your first scan

This is where Guardian starts showing you things you didn't already know.

New CIs are scanned when they connect. To scan again, open the CI from **All CIs** and select **Run Scan Now**, or tick several CIs and select **Scan Selected**. To scan on a schedule, see [section 8](#s8).

### The CI's page

The top of each CI's page shows whether it's online, when it was last scanned and its environment, then a summary: the compliance score and the changes and issues found. Below that, the tabs hold the detail. [CIs](https://docs.cloudhouse.com/guardian/essentials/cis/index.md) explains each tab.

![A CI's page](https://docs.cloudhouse.com/guardian/essentials/assets/screenshots/ci-detail.png)

### Check the scan is complete

Open **Monitored Elements** and look for something you know is there: a service you're sure is running, or a package you know is installed. If Guardian found it, the scan is reading the CI properly.

### If a scan fails or looks incomplete

A failed scan is almost always connectivity, credentials, or an agent that has stopped reporting. Go back to the checklist in [section 4](#s4). [Scans](https://docs.cloudhouse.com/guardian/essentials/scans/index.md) explains each scan state.

A scan that **succeeds but returns less than you expected** needs more care. With agentless scanning it usually means the account could connect but couldn't read everything. The result looks like a complete scan, but parts are missing, and Guardian can't report changes in parts it never saw. Fix the account's permissions and scan again.

## 8. Detecting drift

Each time Guardian scans a CI, it compares the result with the previous scan and records every difference. Those differences are the drift.

### Make a test change

With only one scan, there's nothing to compare yet. Rather than waiting for the next scheduled scan, make a small change yourself.

1. Use a test CI, not a production server.
2. Make a small change you can easily undo and that nothing depends on. For example: add a local group, create a user account, install a small package, or stop a service you don't need.
3. Select **Run Scan Now** on the CI.
4. Look at what changed since the last scan on the CI's **Overview** tab.
5. Undo the change, scan again, and you'll see it change back.

You'll see what a change you made yourself looks like in Guardian, which helps when you're looking at changes you didn't make.

### Scheduled scans

**Scan Configuration** controls when scans run, per CI group or environment: daily, weekly, or only when someone asks. A CI can also override its group's schedule from its own page. See [Scan schedules](https://docs.cloudhouse.com/guardian/essentials/scans/schedules/index.md).

Daily scanning is a good starting point. Scanning too rarely means changes build up into large batches that are harder to explain.

### The Detected Drift page

**Detected Drift** lists every change across your estate: what changed (added, modified or removed), on which CI, whether it's matched to a change request, whether it's reconciled, and when it was detected. Open a change to see the value before and after and its effect on the CI's score. Finding out what changed on a server becomes a quick lookup instead of an investigation. See [Detected Drift](https://docs.cloudhouse.com/guardian/essentials/detected-drift/index.md).

![Detected Drift](https://docs.cloudhouse.com/guardian/essentials/assets/screenshots/detected-drift.png)

- **Expect some noise at first.** Laptops and servers on changing networks pick up things like IP addresses and routes that change on their own.
- **Check when it changed.** The timing often tells you who made the change.
- **One action can produce several entries.** A group of related changes on the same CI at the same time is often one event.

Guardian shows you what changed and when. It doesn't know who made the change or why. You record that when you review it, in [section 10](#s10).

## 9. Comparing scans and CIs

Guardian can compare in two ways, each suited to a different question.

| Your question | Comparison | Where to find it |
| --- | --- | --- |
| What changed on this server, and when? | Two scans of the same CI | **Compare with previous scan** on the CI's page, or two scans from its **History** |
| Why does one of these behave differently? | Two different CIs | Tick two CIs on **All CIs** and select **Compare** |

Comparing scans is useful when something started failing on a particular date and you want to see how the server looked before and after. Comparing CIs helps when two servers should be identical but one isn't behaving. Doing that by hand is slow, and the cause is often a single line in a long configuration. Some differences are expected, such as hostnames, IP addresses and unique identifiers. If the comparison shows no meaningful differences, that's useful too: the problem probably isn't configuration.

**When something breaks,** start by comparing the affected CI's latest scan with an earlier one, in case it changed recently. If it hasn't, compare it with a working CI of the same type.

## 10. Was that change authorised?

Guardian has found a change on a server overnight. The next question is whether anyone was supposed to make it.

Guardian can't answer that by itself, because it sees configuration, not intent. Once you've found out, you record the answer in Guardian so nobody has to ask again.

Every detected change starts as **unreconciled**. When you approve it, with a reason or a change request reference, it's **reconciled**, and the approval goes into the audit trail with who approved it and why. The Dashboard's **Unreconciled drift** panel and its **Needs your attention** list show what's still waiting.

1. Open **Detected Drift** and filter to **Unreconciled**.
2. Open a change and check it against what you know: a change ticket, the team responsible, or your own work.
3. Select **Approve** if it was legitimate, or **Skip** to move on and come back to it later. See [Approve drift](https://docs.cloudhouse.com/guardian/essentials/detected-drift/approve-drift/index.md).

![Approving drift](https://docs.cloudhouse.com/guardian/essentials/assets/screenshots/approve-dialog.png)

Whatever is left unreconciled is the list to look at closely: changes nobody can explain. On most estates the large majority of changes are routine, so this narrows a long list down to a few that need investigating.

### Reviewing changes yourself

At the top of **Detected Drift**, change reconciliation shows **Manual review**: you review and approve changes yourself. Automatic matching of changes to approved change requests is available as a bolt-on for Guardian Essentials or as part of our Guardian Professional offering. See [Change reconciliation](https://docs.cloudhouse.com/guardian/essentials/change-reconciliation/index.md).

It works best when someone does it regularly:

- **Review weekly.** A short weekly review is easy while the changes are recent. Leave it a month and the list becomes too long, and people stop reviewing it.
- **Say why.** Give a reason or the change ticket when you approve. "Approved" on its own won't mean much in six months.

## 11. Proving compliance

An auditor has asked for evidence that your servers meet a recognised hardening standard. This section covers how to produce it.

Guardian assesses CIs against **policies**, on every scan, and **benchmarks** such as CIS Benchmarks, on the benchmark's own schedule. Both are assigned to CI groups. See [Compliance](https://docs.cloudhouse.com/guardian/essentials/compliance/index.md).

### Assign a benchmark

On **Benchmarks**, search for your operating system, open the benchmark and assign it to the CI groups to assess. Results appear once the benchmark's schedule has run. See [Benchmarks](https://docs.cloudhouse.com/guardian/essentials/compliance/benchmarks/index.md).

![Benchmarks](https://docs.cloudhouse.com/guardian/essentials/assets/screenshots/benchmarks.png)

**Choose the level carefully.** CIS benchmarks come in levels. Level 1 is a set of broadly safe recommendations, sensible for most systems without getting in the way. Higher levels are stricter and more likely to affect how systems run. If you're not sure, start with Level 1.

### Read the results

**Scan Results** shows the failed checks across your estate by severity, with each check's current and expected value. Each CI's page shows its own failing policies and benchmarks.

![Scan Results](https://docs.cloudhouse.com/guardian/essentials/assets/screenshots/scan-results.png)

!!! note "Expect your first results to show a lot of failures"

    A general-purpose server assessed against a hardening standard for the first time usually fails many rules. It doesn't mean your estate is in poor shape. Nearly every first assessment looks like this.

1. **Read the failures before you fix anything.** Some will be real gaps. Others won't apply to how you use the server.
2. **Decide what you'll act on.** Reaching 100% is rarely a good use of time. Focus on what matters for your risk and your auditor.
3. **Fix, rescan and check the score improves.**

A benchmark's page shows how its results trend over time. A trend is often more convincing to an auditor than a single figure, so start measuring before you start fixing things.

### Export evidence

**Reports** › **Compliance** shows passed and failed checks per policy, scoped by environment, CI group, policy and severity. See [Reports](https://docs.cloudhouse.com/guardian/essentials/compliance/reports/index.md). Export at the time you're asked, and file it with your other evidence. A report from three months ago describes your servers as they were three months ago.

## 12. Going further with Guardian

If you've worked through this guide, you now have CIs scanned on a schedule, changes being detected and reviewed, and policies and benchmarks reporting on compliance. That's a strong foundation for governing change across your estate.

### Growing with Guardian

As your needs grow, these capabilities are available as a bolt-on for Guardian Essentials or as part of our Guardian Professional offering:

- **Automatic change reconciliation:** detected changes matched to approved change requests in your service management tool, so routine change is reconciled for you ([section 10](#s10)).
- **Tailored policies:** create and manage policies written for your own environment, alongside Guardian's best-practice policies and CIS Benchmarks.
- **Third-party credential stores** such as Delinea Secret Server, alongside Guardian's built-in vault.
- **Risk scoring and end-of-life software tracking.**
- **File integrity monitoring.**
- **AI-assisted change control.**

Your Cloudhouse contact can help you decide which fits your next step.

### Keeping it working

1. **Review unreconciled changes every week.** This is the habit that slips first, and section 10 depends on it.
2. **Keep an eye on CI status.** A CI that's offline or failing scans isn't being monitored, even though it's still in the list.
3. **Add new servers as you build them.** Guardian can't report on a CI it doesn't know about.

### A sensible next step

So far you've worked with a test CI and a first import. Expand gradually. Choose the group of systems where an unexpected change would cause the most trouble, bring those in, assign a benchmark, and review their changes weekly for a month. Then add the next group.

A smaller set of servers that someone reviews regularly is worth more than a whole estate that nobody looks at.

## Getting help, and telling us what didn't work

We'd like to hear where this guide fell short. The most useful feedback is small and specific: where you had to guess, a screen you couldn't find, a step that assumed something you didn't have, or a term we used without explaining it. Use **Was this page helpful?** at the bottom of this page, or tell your Cloudhouse contact. If you got stuck and worked it out yourself, we'd like to hear about that too.

To ask a question or report a problem, email [support@cloudhouse.com](mailto:support@cloudhouse.com).

- [*Learn* **Guardian concepts** The words Guardian uses, and what they mean.](https://docs.cloudhouse.com/guardian/essentials/get-started/concepts/index.md)
- [*Set up* **Groups and environments** Organise CIs so schedules and policies apply to the right ones.](https://docs.cloudhouse.com/guardian/essentials/cis/groups-and-environments/index.md)
- [*Check* **Compliance** Assess CIs against policies and benchmarks.](https://docs.cloudhouse.com/guardian/essentials/compliance/index.md)
