EDR, Antivirus & Backups
Your JD asks for "Knowledge of backup solutions and antivirus/EDR tools". This is a ticket-driven topic as much as a technical one.
AV vs EDR
| Traditional AV | EDR | |
|---|---|---|
| Detection | Signatures of known malware | Signatures plus behaviour and telemetry |
| Scope | Files in real time | Whole endpoint, continuously |
| Investigation | Limited — scan and hope | Full timeline, process tree, network connections |
| Examples | Defender for Antivirus, Trend Micro AV | Defender for Endpoint, CrowdStrike, SentinelOne |
Prevention blocks execution outright. The user gets "this app was blocked" and cannot run it. Detection allows execution, then flags it. The user gets a quarantine notice and the file disappears later. Same root cause, completely different ticket, completely different conversation with the user.
Triage workflow
Get-MpComputerStatus |
Select-Object AMServiceEnabled, AntivirusEnabled, RealTimeProtectionEnabled
Get-MpThreatDetection |
Select-Object InitialDetectionTime, ThreatID, ActionSuccess, Resources
Get-MpThreat | Select-Object ThreatName, SeverityID, CategoryID
| Field | Read it for |
|---|---|
ActionSuccess | Did it actually contain it? |
Resources | The exact file path — needed before any exclusion |
InitialDetectionTime | Correlation with what the user was doing |
SeverityID | High and critical get escalated, not excluded |
Handling "my app is blocked"
The correct sequence:
- What exactly is blocked? Full path, detection name, detection type.
- Is it prevention or detection? Determines whether the file still exists.
- Is it a known false positive? Check the vendor's exclusion guidance.
- Is the file signed and from a known vendor? Trust anchor.
- Get approval, then exclude — minimally.
# Restore a quarantined file
Get-MpThreatDetection | Where-Object {$_.InitialDetectionTime -gt (Get-Date).AddHours(-2)} |
Select-Object ThreatID, Resources
# Define a narrow, explicit exclusion — never a whole drive or extension
Add-MpPreference -ExclusionPath "C:\Program Files\Vendor\App\app.exe"
Remove-MpPreference -ExclusionPath "C:\Program Files\Vendor\App\app.exe"
Excluding C:\, an entire Program Files tree, or a file extension disables
protection for everything that shares that path. That is how ransomware
arrives unopposed. Exclude one file, in writing, with an approval and a review
date.
Defender essentials
# Cloud-delivered protection
Set-MpPreference -MAPSReporting 2
Set-MpPreference -SubmitSamplesConsent 1
# Exclusions — what is currently set?
Get-MpPreference | Select-Object -ExpandProperty ExclusionPath
# Scan history and a manual scan
Start-MpScan -ScanType QuickScan
Start-MpScan -ScanType FullScan
Start-MpWDOScan # offline scan — catches rootkits
| Concept | Why |
|---|---|
| Real-time protection | Should never be disabled to "fix" a problem |
| Tamper protection | Prevents users disabling Defender |
| Cloud-delivered (MAPS) | Uses live reputation data |
| Offline scan | Boots without drivers — finds rootkits |
Turning off real-time protection to install software, and never turning it back on, is one of the most common and most damaging things a new support engineer does. If it must be temporarily disabled, write down that you did it and re-enable it immediately after.
Backups
| Type | Stores | Restore needs |
|---|---|---|
| Full | Everything | One set |
| Incremental | Changes since the last backup of any type | The entire chain |
| Differential | Changes since the last full | Full plus one |
RPO and RTO
- RPO — how much data loss you can tolerate. Measured in time.
- RTO — how long you can be down. Measured in time.
| Workload | Example RPO / RTO |
|---|---|
| File server | 24 h / 8 h |
| Database | 15 min / 1 h |
| Domain controller | 1 h / 4 h |
3-2-1
3 copies, 2 media types, 1 offsite.
# Windows Server Backup
Get-WinEvent -LogName 'Microsoft-Windows-Backup/Operational' -MaxEvents 10
wbadmin get versions
| Tool | Notes |
|---|---|
| Windows Server Backup | Built in, adequate for small shops |
| Veeam / Acronis | Industry standard for VMs |
| Azure Backup | Offsite by definition |
| Restic / Borg | For Linux, deduplicated |
| rsync + cron | Linux servers, simple and effective |
A green "backup completed" log proves the job ran. It proves nothing about restorability. Schedule periodic restore tests, and know the last one by date — that is the question an auditor asks first.
Backup failures you will meet
| Symptom | Cause |
|---|---|
| "There is not enough space" | Destination full, or VSS snapshot reserve |
| "The backup could not be completed" | A file in use, or permissions |
| Free space drops during backup | Shadow copy storage reserve too small |
| Backup runs forever | The source is on slow storage, or over a WAN |
| Last restore test failed | The backup was never valid |
Interview scenario
'My business app was blocked by antivirus' — how do you answer?
"I'd first establish whether it's prevention or detection, because that decides whether the user can still open the application. Then I'd get the exact file path and detection name from the console — never add an exclusion from a paraphrase.
Then I'd check whether it's a recognised false positive for that vendor. If it is, I'd raise a change with the file owner, exclude that specific file path rather than a folder or extension, record the approval, and set a review date. If it isn't a known false positive, I'd treat it as a potential compromise and escalate rather than silence it."