How long to retain SOC logs? Balancing investigation, compliance, and costs
Should enterprises retain SOC logs for 30 days, 90 days, 180 days, or 12 months? This is one of the critical questions when building a security operations center or evaluating outsourced SOC services.
There is no single optimal SOC log retention period for every enterprise. The appropriate timeframe depends on the type of security logs, incident detection and investigation capabilities, applicable legal requirements, data retrieval speed, and operational budget.
In practice, enterprises can design multi-tiered retention policies: data requiring frequent analysis is kept in a fast query tier, while rarely used historical logs are moved to a lower-cost storage repository. However, log retention time does not equal the time data is directly searchable on the SIEM system.
Therefore, before choosing a SOC package with 90 days of searchable logs or a 12-month archive, enterprises must clearly understand the purpose of each timeframe and the ability to exploit data when an incident occurs.
Why shouldn't SOC log retention time be a fixed number?
Logs are data recording events occurring on IT systems, from account logins and access permission changes to network connections and server activities. In the operations of a security operations center (SOC), logs serve to provide evidence for anomaly detection, alert analysis, and incident investigation.
However, not all types of logs hold the same value or retention requirements.
For instance, privileged account authentication logs can be useful when investigating prolonged unauthorized access behaviors. Meanwhile, some operational logs with high frequency but lower security value may not need to be maintained in a high-performance storage tier for the same period.
According to the Best Practices for Event Logging and Threat Detection guidelines, published in August 2024 with the participation of multiple international cybersecurity agencies, a logging policy must define the type of events recorded, monitoring mechanisms, retention periods, and policy review intervals.
There are four main objectives when determining SOC log retention times.
Threat detection. The SOC needs near real-time data to correlate events and identify suspicious behaviors. For example, multiple failed logins followed by a successful login from an unusual location.
Incident investigation. An incident might only be discovered after the intrusion has occurred. If relevant logs are deleted before the investigation, the security team will struggle to identify compromised accounts, the chain of actions, or the scope of damage.
Audit and compliance. Certain systems or organizations may be required to maintain log data according to legal regulations, industry requirements, applicable standards, or contractual obligations.
Optimizing operational costs. Storing all logs in the fast query tier for extended periods can significantly increase SIEM, storage, and data management costs.
Therefore, selecting retention is not simply a decision to purchase more storage capacity. It is a decision about the level of investigation readiness and security capability the enterprise wishes to maintain.
Which types of logs should be prioritized for longer retention?
A common mistake in designing SIEM log retention is applying the same timeframe to all data sources.
This approach has two limitations. If the timeframe is too short, the enterprise may lose critical data. Conversely, if all logs are kept for a long time, costs may increase while the investigative value does not increase proportionally.
Security logging policies should begin by classifying log sources according to assets, risk levels, and scenarios requiring investigation.

Identity and Authentication Logs
This group includes login logs, multi-factor authentication (MFA), password changes, permission grants, and administrative account activities.
This is a crucial data source when analyzing account takeover behaviors, unauthorized access, or privilege abuse.
Enterprises using Active Directory, Microsoft Entra ID, or centralized identity management systems need to consider the ability to correlate authentication logs with endpoint, VPN, and cloud data.
Firewall, VPN and Network Logs
Firewall, VPN, and network device logs assist in tracing connections, source addresses, destinations, and access times.
However, storing all traffic logs at high granularity can generate massive data volumes. The SOC should distinguish logs used for threat detection from network data needed solely for routine operational purposes.
Endpoint and EDR Logs
Endpoint data provides information about processes, execution behaviors, system changes, or signs of anomalous activities.
Events with high investigative value must be prioritized for retention. Enterprises must also distinguish data exported to the SIEM from the telemetry history kept within the EDR/XDR platform itself, as the retention times of the two systems may differ.
Cloud Audit and Critical Server Logs
Cloud logs record administrative activities, configuration changes, resource access, and operations with cloud services.
For critical servers and databases, events such as permission changes, access to sensitive data, or modifications to security configurations need to be considered in the retention policy.
In general, the types of logs that need to be retained longer are those supporting a specific investigation requirement or retention obligation, not necessarily those with the largest volume or generated most frequently.
Enterprises can further refer to the SOC architecture with Log, SIEM, SOAR, and EDR/XDR to understand the role of each data source in the monitoring system.
Designing SOC log retention times using Hot, Warm, and Cold Storage
One method to balance investigative effectiveness and costs is applying a multi-tiered storage model.
Instead of maintaining all logs in the SIEM with the same query performance, data can be allocated based on usage frequency and required retrieval times.
Hot Storage: Data for monitoring and rapid investigation
Hot storage is the storage tier prioritizing fast search and analysis capabilities, typically used for newly occurring events.
This is the data the SOC needs to query frequently to verify alerts, correlate events, and analyze suspicious activities.
The cost of hot storage is usually higher than less accessed storage tiers, especially when platforms charge based on ingested data, indexed data, or searchable capacity.
Warm Storage: Data for extended investigation
Warm storage is suitable for data no longer queried frequently but still requiring search or historical analysis capabilities with acceptable latency.
For example, upon detecting a compromised account, SOC analysts may need to check the account's activities over the preceding months.
The search capability on warm storage depends on the deployed architecture and platform; it is not inherently identical to hot storage.
Cold Storage: Long-term historical archiving
Cold storage or archive is used for infrequently accessed data that still needs to be preserved for investigation, auditing, or other requirements.
The advantage is the ability to reduce storage costs per data unit. In return, enterprises may have to accept longer restoration times or additional costs upon retrieval.
An important point is that archived data should not be viewed as data always ready for instant searching.
Retention Matrix: A reference SOC log retention model
The table below is an illustrative design framework for enterprises building retention policies. The timeframes are technical assumptions for planning, not legal minimums or mandatory standard periods for SOC services.
Log type | Hot | Warm | Cold / Archive | Priority objective |
Identity / Authentication | 30 days | Next 60 days | Per policy | Account takeover investigation |
Firewall / VPN | 14–30 days | Up to 90 days | Per policy | Connection tracing |
Endpoint / EDR | 30 days | Up to 90 days | Per policy | Endpoint behavior investigation |
Cloud Audit | 30 days | Up to 180 days | Per policy | Cloud activity tracing |
Critical Server / Database | 30 days | Up to 180 days | Per policy | Privileged access investigation |
Security Alerts / Cases | 90 days | Up to 180 days | Per policy | Incident record cross-referencing |
*Note: Hot and warm intervals represent consecutive storage phases and are not cumulative if marked "up to". Cold/archive can follow warm; the total retention period must be determined and approved separately.
For example, if an enterprise determines it needs to keep a log group for 12 months, a reference configuration could be 30 days hot, the next 60 days warm, and the remaining 275 days cold, assuming a 365-day year.
However, this approach is only reasonable if the restoration speed from cold storage meets investigation requirements, while the storage method complies with all applicable obligations.
Legal hold is an exception that needs to be designed from the start. When logs relate to a dispute, investigation, or evidence preservation request, enterprises may need to pause the automated deletion process for the exact relevant dataset, according to approved procedures.
Does keeping logs longer mean better investigation capabilities?
Not necessarily.
An enterprise might retain logs for 12 months but still face difficulties when investigating incidents if the data is unsearchable, lacks context, or loses integrity.
This is why three concepts must be distinguished when evaluating log retention capabilities:
Retention Period: The total time data is retained.
Searchable Retention: The timeframe during which data can be queried on the analysis system.
Retrieval Time: The time required to restore data from the storage tier or archive.
For example, a SOC service package offering 90 days of searchable retention and 365 days of archive does not mean the SOC can instantly search the entire 365 days of data.
When an incident from six months prior arises, the investigation team may have to request data restoration, reload it into the analysis environment, and perform event correlation.
Besides retrieval time, log quality also depends on other technical factors.
Time synchronization: Devices must use a reliable time source and standardize time zones to avoid discrepancies when reconstructing the incident timeline.
Completeness: Logs need to record fields appropriate for the investigation objectives, such as user identities, assets, timestamps, and action types.
Integrity: Measures must exist to prevent unauthorized modification or deletion. Mechanisms such as access control, encryption, immutable storage where appropriate, and integrity checks help increase data reliability.
Correlation capability: Logs from multiple sources must have sufficient linking information for SOC analysts to reconstruct the sequence of events.
The NIST SP 800-92 – Guide to Computer Security Log Management guidelines emphasize building and maintaining a unified log management process across the entire organization. This is a foundational document on security log management, although enterprises still need to cross-reference with current technologies and regulations.
For enterprises evaluating SOC services, the question to ask is not only "how long are logs retained?" but also "when an investigation is needed, what data can we retrieve and how quickly?".
This need relates directly to SOC incident investigation processes and skills, where log data acts as the foundation for analyzing attack indicators.
How are SIEM log retention costs calculated?
SOC log retention costs do not depend solely on the number of storage months. Two enterprises requiring 12-month retention might incur vastly different budgets due to varying daily log generation volumes, numbers of data sources, indexing methods, and storage architectures.
When calculating the Total Cost of Ownership (TCO), enterprises need to consider these key components:
Ingestion: Costs of receiving and processing logs from firewalls, servers, endpoints, cloud, and other sources.
Indexing and Search: Resources used for indexing, correlating, and querying data.
Storage: Capacity required to maintain the hot, warm, and cold tiers.
Retrieval: Costs to restore, read, or transfer data from the archive.
Operations: Personnel, maintenance, pipeline monitoring, and retention policy management.
An initial capacity estimation method is:
Log capacity ≈ GB/day x Number of retention days |
For example, an enterprise ingests an average of 100 GB of logs per day. If all raw data is retained for 365 days, the theoretical capacity would be:
100 x 365 = 36.500 GB |
Equivalent to approximately 36.5 TB in decimal units, before accounting for compression, data replication, indexing, and other factors.
This is merely a capacity calculation, not an actual cost estimate. Notably, some SIEM platforms charge by ingest volume, while other solutions charge based on retention time, features, or query resources.
How to optimize costs without losing critical data?
Enterprises can apply three principles.
First, prioritize log collection based on security use cases. Not all operational events need to be fed into the SIEM. However, rigorous evaluation is required before filtering or discarding data to avoid creating investigation blind spots.
Second, separate searchable time and archive time. Less frequently queried log groups can be moved to lower-cost storage tiers as long as they still meet retrieval requirements.
Third, periodically reassess log volumes. Adding new cloud systems, endpoints, or applications can significantly increase ingest volumes compared to initial estimates.
When comparing an internal SOC with an outsourced SOC, enterprises also need to account for infrastructure costs, SIEM licensing, personnel, and system administration. These factors are analyzed more deeply in the article on outsourcing SOC and the security budget optimization problem.
The important thing is not reducing log retention time at all costs, but choosing an option where the cost is commensurate with the capabilities for detection, investigation, and meeting retention obligations.
What legal requirements must SOC log retention periods meet?
For enterprises operating in Vietnam, log retention policies must be evaluated alongside regulations on cybersecurity, information security, personal data protection, and relevant industry-specific requirements.
However, one should not assume that every enterprise in Vietnam must retain all SOC logs for the same period.
According to information from the Government Electronic Information Portal, Cybersecurity Law No. 116/2025/QH15 takes effect on July 1, 2026. Furthermore, Decree 333/2026/NĐ-CP, detailing a number of articles and measures to implement the Cybersecurity Law, takes effect on August 19, 2026.
These documents form part of the basis for organizations to review their cybersecurity protection responsibilities. But determining specific log retention obligations still requires cross-referencing applicable subjects, system types, data types, and individual relevant clauses.
Enterprises should implement three steps when building a retention policy.
Step 1: Define the scope of application. Classify information systems, operational models, processed data, and related legal or contractual requirements.
Step 2: Cross-reference with retention requirements. Determine which types of logs must be preserved, for how long, requirements for data integrity, and the ability to provide them upon valid request.
Step 3: Approve policies and exceptions. The information security department coordinates with IT, legal, and relevant units to agree on retention times, access rights, deletion mechanisms, and review procedures.
Legal hold and evidence preservation requests
Legal hold is a mechanism that suspends standard scheduled data deletion or destruction activities when the information is related to a dispute, investigation, or evidence preservation request.
For instance, a log group nearing the end of its retention period but currently serving an incident investigation may need to be retained based on an authoritative decision within the organization.
This procedure should clearly identify the approver, log scope, start time, termination conditions, and how data is protected during the hold period.
Additionally, logs may contain IP addresses, account identifiers, or information potentially related to individuals. Therefore, retention periods and purposes must be considered alongside data protection requirements, avoiding the collection and retention of information beyond legitimate needs.
Enterprises can further consult the context on cybersecurity and data retention regulations in Vietnam, while directly cross-referencing current legal documents.
Note: The content above provides a technical and administrative evaluation framework and does not substitute for legal advisory opinions. The retention periods in this article are design examples, not affirmations of universal obligations applicable to all organizations.
Before outsourcing SOC services, what do enterprises need to confirm regarding log retention?
When evaluating a SOC provider, enterprises often focus on 24/7 monitoring capabilities, SIEM technologies, or alert response speeds.
However, log retention time and retrieval capabilities must also be considered crucial criteria within the service scope.
A SOC package advertising a 12-month retention time is insufficient to conclude that an enterprise will have all necessary data for every investigation.
Before signing a contract, enterprises should clarify:
Log source scope: Which systems are collected, what data is ingested into the SIEM, and are there capacity limits?
Query time: How long are logs directly searchable? Does archived data require restoration before analysis?
Retrieval SLA: When requesting historical logs, what is the provider's committed processing time?
Additional costs: How are fees calculated for increased ingestion, retention expansion, data exports, or restoration?
Data ownership and security: Who has access to the logs, where is the data stored, and what protective controls are in place?
Handover upon contract termination: In what formats can the enterprise export data, and how is the post-handover data deletion process executed?
Legal hold: Does the provider support preserving logs related to incidents upon request and under approved procedures?
These points must be reflected in the service scope, data management policies, and responsibility agreements between the parties.
In particular, enterprises must distinguish between SOC monitoring services and Managed Detection and Response (MDR) services, as the scope of telemetry collection, analysis, and response can differ by contract. The article How SOC and MDR differ helps clarify this distinction.
For a collaborative model between internal IT and an MSSP, the retention policy must also establish who manages log sources, who approves changes, who performs restorations, and who is responsible for handling data provision requests. This is also relevant to the Co-managed SOC model and how to divide responsibilities.
A suitable SOC contract determines not just log retention times, but must also clarify the usability of the data when the enterprise genuinely needs to investigate.
Conclusion: How long is it reasonable to retain SOC logs?
The SOC log retention period should be decided based on the security objectives, investigation requirements, applicable obligations, and operational budget of each enterprise.
The 30, 90, 180-day, or 12-month milestones can serve as reference options when designing a retention matrix, but should not be viewed as universal standards.
An effective policy must clearly define which log types must be kept, searchable time, archive time, restoration capabilities, data integrity, and data management responsibilities.
For enterprises considering outsourcing a SOC, the value to evaluate is not merely the number of months logs are kept, but the ability to detect, retrieve, and correctly utilize the right data during an incident investigation, at an appropriate operational cost.
SOC 24/7 solution from IPSIP Vietnam
IPSIP Vietnam provides 24/7 SOC services combining a team of experts with security monitoring and analysis technologies. Enterprises can refer to IPSIP's 24/7 SOC services and book a consultation to discuss monitoring scopes, log sources, retention needs, and deployment options suitable for their existing infrastructure.

References:
1. NIST SP 800-92 – Guide to Computer Security Log Management
2. Australian Cyber Security Centre – Best Practices for Event Logging and Threat Detection
3. Government Electronic Information Portal – Cybersecurity Law No. 116/2025/QH15
4. Government Electronic Information Portal – Decree 333/2026/NĐ-CP










