Making security event logs useful
Choose useful security events, check collection and timestamps, protect log access and assign a tested response process for alerts and investigations.

Security logs record selected activity in systems and services. They can help investigate an incident, but only if the needed events were collected, retained and available to the people doing the investigation.
Start with questions you need to answer: who changed an administrator role, whether a sign-in succeeded, or which account changed a sharing setting. Then check that the chosen system records enough evidence.
Decide which events matter
Identify critical accounts and services with IT. Prioritize sign-in outcomes, privilege changes, security-setting changes and relevant access to sensitive information. Record what each source captures and what it leaves out.
The joint event-logging guidance endorsed by the Cyber Centre covers policy, centralized collection, integrity and detection. Apply it to your environment’s needs and capabilities.
Don’t assume every cloud plan includes the same event types or retention. Confirm what your licence includes, which settings are on and when collection started. Logs enabled today won’t recreate yesterday’s missing events.
Check collection with known actions
Use a test account to perform a few harmless actions agreed with IT. Confirm that the record shows the correct identity, outcome, source and timestamp. Check failed actions as well as successful ones.
Central collection can make related events easier to investigate. Verify the connection from each source and alert when expected records stop arriving. Include clock synchronization and time-zone interpretation in the checks.
Keep the original event alongside any normalized fields where practical. A parser that drops an important detail can make a central search misleading.
Protect records and their administration
Limit access to logs, export tools and retention settings. Separate administration where feasible so compromising a monitored account doesn’t also provide control over all its evidence.
Protect records in transit and storage. Encryption protects confidentiality under its particular key and access arrangements, but it won’t stop records being deleted or altered.
If you use immutable storage, verify its configured retention and administrative boundaries. Describe the protection you’ve tested, and don’t call the records tamper-proof.
Set retention for a documented purpose
Balance investigation needs, contractual and applicable legal requirements, privacy and cost. Logs may include personal information or confidential operational details. Avoid recording passwords, tokens or unnecessary message content.
Document how long each source and central copy lasts, who signs off exceptions and what happens during a preservation request. Check export and retrieval before an investigation depends on them.
Assign and test alert handling
Choose alerts that someone can investigate. State who receives them, their coverage hours and the next contact if they don’t respond. A tool can keep sending alerts when no one is there to read them.
Run a harmless test event and follow it through notification, assessment and the work record. Verify that the responder can retrieve supporting logs. Review noisy alerts and missed collection separately.
The Cyber Centre’s SIEM guidance can help assess whether a centralized analysis platform fits. Start with the coverage and response you need, then choose the tool.


