Skip to content

Latest commit

 

History

History
190 lines (153 loc) · 6.29 KB

incident-response.md

File metadata and controls

190 lines (153 loc) · 6.29 KB

Incident Response Topics

Semantics

Types of Incidents

  • Data leak
  • Breach / penetration
  • Leak/Breach from Partner
  • DDOS
  • Fraud
  • Platform SLA
  • Partner-specific SLA

Timeline elements for Reporting

  • incident start
  • incident end
  • investigation stage
    • system scope
    • product / LOB scope
    • user scope
    • partner scope
    • client scope
  • escalation/decisions/changes stage
  • Customer impact area(s). Evidences.
  • First client comms
  • each ongoing client comm
  • client followup comms

Breach Announcement

As with all matters related to data breach, research should be kept confidential and shared only with senior leadership involved in the announcing the breach.

Time periods

  • Of downstream breach
  • Of overlap

Totals

People
  • those with PII leaks
  • in each class of data
Types of data revealed
  • particularly high-status data
    • PII, PCI, PHI
    • specific fields

Accounts, passwords, SSN, address. I think this is probably the data classes that we recently defined in the RFC for data classes.

  • Financials, if money stolen.

  • Scope of financial burden for remediation.

  • Special topic. Not always needed.

  • Partners that were involved

    • Follow executive guidance on Partner types / classification.

Degree of certainty

about any of these figures, particularly the number of people impacted.

  • No more that
  • between X-Y
  • was exactly
  • likely was less

Incident Scopes

Priority Kinds

Priority #1 - Emergency/Urgent:

An Incident has caused a complete and immediate work stoppage affecting at least one primary business process or a broad group of End Users. No workaround is available.

Priority #2 - High:

An Incident has affected a business process in such a way that business functions (operations) are severely degraded, multiple End Users are impacted or key external customer is affected. A workaround may be available, but it is not easily sustainable.

Priority #3 - Medium:

An Incident has affected a business process in such a way that certain business functions are not available to End Users or external customers, or a system or service is degraded. A workaround may be available.

Priority #4 - Low:

An Incident has little or no effect on business processes or operations and can be handled on a scheduled basis (e.g., preventive maintenance). A workaround is available.

Incident will then continue with containment, eradication and post incident analysis. Bi-Yearly tabletops are held to test the program

SLA: definitions

  • In contracts / ensure standard language
    • impact times
    • contact info during incident (us & them)
    • All contracts / SF records ideally to be updated
  • Account annotation for modification from standard
    • Zak put this in Salesforce for BD/PS already
  • Ops-driven Dashboards / ‘pager’ alerts
  • Runbooks for each impact-level (amber, red)
    • error codes, (for app colophons)
    • Comms templates, internal
    • Comms templates, client (perhaps: major, vs. minor)
    • Escalation triggers
    • Logging

Data breach, internal

  • Contract language for data breach contact
  • Ops-driven Dashboards / ‘pager’ alerts
    • IDS (intrusion)
    • PII / PHI / PCI movement
  • Runbooks
    • Forensics / log preservation (may need training)
    • Retained cyber analyst, vector analysis
    • Warroom
    • Impact scope
    • Comms templates, affected partners (supply & demand)
    • Comms templates, unaffected partners (supply & demand)
    • Comms templates, press
      • include Common elements
    • Comms templates, board/investor

Data breach, partner

  • Same as internal, plus:
  • establish contact with partner sec team, co-warroom
  • Timeline analysis pinpoint Even exposure window
  • Analyze scope (specific fields of PII, PCI, PHI)
  • Comms templates, internal
  • Comms templates, unaffected partners (supply & demand)
  • Comms templates, press
    • include Common elements
  • Comms templates, board/investor

DDOS

  • Runbook for extortion / threat-comms
    • capability assessment
    • cyberSec / DDOS contractor retained
    • comms control
      • Warroom
      • Negotiation / delay
    • FBI escalation threshold
  • Runbook for live attack
    • Activate static and API caching
    • Global traffic response / readiness
    • Comms template: partner, SLA/SLO notification bluff
    • Comms template: partner transparency

Timetable for (minimum) yearly exercises.

  • Needs to be in corp cal
  • Retro / minutes from each exercise
  • New/old staff

Fraud

  • Same as internal, plus:
  • legal review of related insurance policies
    • operationalize following guidance to enable claims
  • if needed, establish contact with related partners
  • Comms templates, board/investor

Per-incident occurrence

  • Incident ‘war room’
    • Generally private, to protect forensic discussion in a data breach
    • Probably use the same warroom (slack channel) for all incidents.
    • At bigger corps, potentially diff channels for each potential incident, for forensic trail/logs
  • Post-mortems
    • Produce tickets for action items
    • Minutes
    • Breaches needs SECCOM ticket with label review-security

Involvements

  • What communications template(s) were used?
  • What medium were the comms sent through? (social media channels, a service desk, a status page, email, etc)
  • Who sent each out?
  • Contact address for where each affected party?

Post mortems

Timeline

  • If we did not communicate early and often with customers, why not?

Involvements

  • What path did we send customers down when they reached the error page?
  • Is this a typical flow for customers when there is an error?
  • What atypical solutions were used or developed with customers?
  • How did we determine who would communicate out to customers?
  • Did we communicate to the right customers? How do we know?

CX

  • How would you have felt as a customer if you encountered a similar issue?
  • Did the comms we sent out reduce the need for customers to contact our support team?
  • Were the mediums/channels used easily accessible/visible for customers? How did they know where to find it?
  • What do our customers now know about this incident solely from the comms we sent out during/after the incident?

Notes

For audits, most of these things will need some kind of screenshot, for proofs

Have identified press / journals, and alternatives