D033F2
Observability
📊 Honey Log Group
Decoy CloudWatch/Azure Monitor log groups.
What it is
A log group named like a security log, with a little plausible content. Attackers delete or shorten logs to hide their tracks, often in bulk. Deleting the decoy log group or changing its retention are logged control-plane events, so log tampering touches the decoy along with real logs.
Implementation requirements
- AWS: a CloudWatch Logs log group named like a security log (audit, GuardDuty export, VPC flow logs) with a little plausible content. Azure: a Log Analytics table or a diagnostic setting with an equivalent name
- Monitoring, tampering: writes that name the decoy (
DeleteLogGroup,PutRetentionPolicy,DeleteSubscriptionFilter, or deleting the Azure diagnostic setting) are logged control-plane events. Defence-evasion tooling that wipes logs touches the decoy along with real groups - Reads (
GetLogEvents,FilterLogEvents) may not reach CloudTrail with default settings. Test before relying on them - Safety: the decoy holds no real log data. Protect real log groups with the same controls, so the decoy is not the only one an attacker can delete
- Rollback: Delete the log group or diagnostic setting.
- Upkeep: Write a little plausible content now and then, so it does not look abandoned. Quarterly proof-of-life retention change from a test principal.
Technologies
AWSAzure
ATT&CK coverage 5
| Technique | Tactic | Status | Confidence | Tier |
|---|---|---|---|---|
| T1685 Disable or Modify Tools |
Defense Impairment | theoretical | moderate | |
| ↳ Detection point: the attack step itself touches the decoy | ||||
↳ Evidence (3 sources)
| ||||
| T1685.002 Disable or Modify Cloud Log |
Defense Impairment | theoretical | moderate | |
| ↳ Detection point: the attack step itself touches the decoy | ||||
↳ Evidence (2 sources)
| ||||
| T1526 Cloud Service Discovery |
Discovery | theoretical | moderate | T1 |
| ↳ Detection point: the attack step itself touches the decoy | ||||
↳ Evidence (3 sources)
Moderate: discovering that a service exists does not necessarily touch a decoy resource in it. | ||||
| T1654 Log Enumeration |
Discovery | theoretical | moderate | T2 |
| ↳ Detection point: the attack step itself touches the decoy | ||||
↳ Evidence (2 sources)
Reading a decoy log is logged only where read access is audited. | ||||
| T1070 Indicator Removal |
Stealth | theoretical | moderate | T2 |
| ↳ Detection point: the attack step itself touches the decoy | ||||
↳ Evidence (2 sources)
Deleting a decoy log group or alarm is a logged management event. | ||||
| ↳ Good fit (tampering with decoy logs is a clean signal); technique evidence is industry-only. | ||||
Evidence
- Decoy Object (D3-DO) — MITRE D3FEND standard
- MITRE Engage — adversary engagement matrix and ATT&CK mapping — MITRE 2022 standard
- MITRE ATT&CK Enterprise v20.0 — MITRE 2026 standard