•

Problem Investigation Intake: Group Incidents Safely

Three similar stone spheres reveal different interiors under a magnifying glass, illustrating evidence-based incident grouping.
What’s in this article?

    Decide which incidents belong in one investigation, preserve contradictory evidence and give the receiving analyst an explicit scope to accept.

    Three tickets say “portal unavailable.” One concerns a browser timeout, another an expired account and the third a request rejected after a release. Grouping them under one problem may look efficient, but a shared phrase is not evidence of a shared cause.

    A problem investigation intake workflow governs the boundary between an incident signal and accepted investigative work. Its work object is an admission case: a proposed scope, a set of incident references, the evidence for grouping them and a decision about who will investigate. The outcome is an accepted new investigation, an accepted link to an existing one, or an explicit return with an owner—not a prematurely declared root cause.

    This is a proposed system design for internal IT-service teams. Adapt its labels and authority to your own problem, incident, access and change policies. It is not a technical troubleshooting procedure or a universal ITIL requirement.

    Keep incident response separate from investigation membership

    In this design, restoration responsibilities and investigative responsibilities remain separate: linking an incident to a problem must not silently change its recovery owner, priority or closure test.

    The incident lead continues response. A problem coordinator manages admission. An analyst assesses the proposed scope and evidence. The service owner resolves scope or resource disputes under local policy. Assign backups for each duty; a named team alone is not an accepted handoff.

    The BMC Helix problem-management process overview describes reviewing incident requests, relating them to an existing investigation when appropriate, or creating an investigation and assigning a specialist with suitable skills, availability and access. The model below makes the evidence and acceptance boundary explicit without assuming another platform implements BMC’s behavior.

    Capture a grouping proposal, not a causal conclusion

    Trigger admission through a recurring pattern, a significant incident needing further investigation, or a service review that identifies a plausible unresolved problem. Define local admission criteria; neither a universal incident count nor a fixed review interval fits every service.

    Give each incident a membership assessment

    Record the service, component if known, environment or version, affected cohort, occurrence window and observed symptom. Link original records and authorized evidence locations. Mark unknowns as unknown rather than filling them with assumptions from another ticket.

    For each proposed member, record the positive evidence, conflicting evidence, current assessment, assessor and review trigger. Suggested assessments are candidate, provisionally linked, supported, excluded and disputed. These labels express investigation membership, not certainty that a root cause has been proved.

    Proposed incident membership decision model
    FindingAdmission decisionRequired evidence or next action
    Similar wording onlyKeep as candidateCoordinator requests distinguishing observations, environment and timeline; do not imply a common cause.
    Relevant conditions align, but cause is unknownProvisional link for investigationAnalyst accepts scope, uncertainty and the observation that would challenge the grouping.
    Existing investigation plausibly covers the same scopeOffer link to its ownerCompare applicability and exclusions; record receiver acceptance before ending the admission case.
    Evidence contradicts the accepted boundaryDispute membershipPreserve prior rationale, name reassessment owner and decide whether to exclude or split.
    Same underlying incident was reported more than onceRecord duplicate-report relationshipKeep distinct report and event counts; retain source references and incident-policy decisions.

    Do not automatically merge original incident records. A link can support investigation while each incident retains its own impact, response history and obligations. Also distinguish duplicate reports of one event from separate recurring events; otherwise a high ticket count can misrepresent recurrence.

    Search existing investigations before creating another

    Compare service, observed behavior, applicability, exclusions and unresolved questions—not just titles. Micro Focus’s investigation and diagnosis process includes validating whether a problem is new or duplicated before continuing analysis. Its workflow is product-specific; the portable principle is to make the duplicate decision visible and attributable.

    A matching known-error article is a lead, not automatic permission to apply its workaround. Check its approved scope through the separate known-error workflow. If its distinguishing conditions do not match, retain the mismatch as evidence.

    Make analyst acceptance the admission completion test

    The coordinator retains progress ownership until the named receiver accepts the packet. Assignment, notification and acceptance are separate events. The analyst can accept a provisional grouping while explicitly rejecting any claim that the cause is already established.

    Admission case states and exit conditions
    StateAccountable dutyCondition to move forward
    Signal capturedCoordinatorIdentify source records, service, reason for investigation and current incident-response owners.
    Evidence pendingCoordinator with evidence suppliersCollect minimum context or document the specific gap and authorized decision needed.
    Scope reviewAnalystAssess membership, exclusions, existing-investigation match and open questions.
    Acceptance pendingCoordinatorNamed investigation owner accepts scope, tasks, access and next checkpoint, or returns explicit gaps.
    Admitted / linkedInvestigation ownerRead back the accepted investigation reference and incident relationships; retain decision evidence.
    Returned / not admittedCoordinatorRecord reason, destination owner, accepted next action and any re-entry condition; incidents remain separately managed.
    Disputed / split reviewAnalyst with service ownerResolve conflicting membership and obtain acceptance for each resulting scope before closing reassignment.

    The handoff packet contains the proposed problem statement, membership register, original evidence references, competing explanations, exclusions, access constraints, priority rationale and investigative questions. Include the first task, target checkpoint and receiver. Do not require a confirmed cause as the entry ticket to work whose purpose is to investigate it.

    The receiver accepts the investigation scope and confirms that required evidence is accessible. The coordinator then verifies the recorded investigation reference and links. An unverified external update leaves acceptance verification open; it is not proof of successful linking.

    Give failed admission paths an owner and a resume condition

    Missing context or inaccessible evidence

    The coordinator names the supplier, missing observation and next checkpoint. Keep the case in Evidence pending. If evidence cannot be shared directly, the authorized evidence owner supplies an approved summary or controlled review path. Escalate unresolved access to the service owner; do not copy sensitive logs into a broader queue. Resume when the analyst accepts the available evidence and its limitations.

    Rejected scope or unavailable investigators

    A rejection must identify the gap: insufficient context, wrong service boundary, unsupported grouping or missing authority. The coordinator returns a specific request, not a generic “needs more information.” A dispute goes to the service owner with both assessments preserved.

    If the analyst lacks capacity, skills or access, use the authorized backup or request a resourcing decision. Keep Acceptance pending until a qualified receiver accepts. Track age and original signal time so reassignment does not make delayed work appear new.

    Duplicates, vendor waits and contradictory results

    For competing investigations, the service owner confirms the authoritative scope with both owners. Transfer unique evidence and obtain receiver acceptance before retiring a duplicate admission case. Do not discard incident references merely because their titles overlap.

    A vendor dependency needs a question, authorized evidence package, supplier contact, internal owner and follow-up checkpoint. The internal owner remains accountable while waiting. A missed checkpoint triggers escalation and a revised plan, not an invented diagnosis.

    When evidence undermines a link, mark membership disputed and pause decisions that depend on the challenged grouping. The analyst proposes exclusion or a separate investigation; the service owner resolves contested ownership. Resume only after the revised scope and destination are accepted and verified. Retain who changed the assessment, why and which downstream readers were informed.

    Worked example: split the symptom without losing the work

    In this constructed example, four portal incident records enter scope review. Two describe timeouts on release A in the same service. One describes an account-expiry message on release A. The fourth reports a timeout on release B, but its diagnostic evidence is not yet available.

    The analyst accepts the first two as a provisional investigation cohort. Their cause remains unknown. The account-expiry case is excluded with a recorded reason and an accepted handoff to its existing incident owner; exclusion does not mean it is resolved. The release-B case remains a candidate while an authorized evidence owner supplies distinguishing observations.

    Later evidence shows a different failure path for release B. The coordinator offers it to a separate investigation owner and verifies acceptance. The original register now accounts for all four records: two provisionally linked, one excluded and returned, one transferred to a separately accepted investigation. No incident was closed by the grouping decision.

    Subsequent causal analysis belongs in the root cause analysis process. Admission has succeeded when the investigative work is owned and its boundary is honest—not when the team has prematurely written “root cause found.”

    Measure admission quality, not just accepted case counts

    • Demand: signals, proposed cohorts and distinct incidents, separated from duplicate reports.
    • Flow: time from signal to accepted ownership, plus time in evidence and acceptance states.
    • Queue health and capacity: oldest unaccepted case, overdue checkpoints and waiting cases by required investigator skill.
    • Quality: returned packets, inaccessible evidence, disputed links and splits caused by overlooked contradictory observations.
    • Reliability: failed external-link verification and unaccepted transfers.
    • Outcome: accepted investigative scopes with traceable membership and next actions; review whether later findings support or overturn admission assumptions.

    Use recurring returns to improve intake questions. Use repeated scope disputes to refine distinguishing checks. Do not infer incidents prevented or hours saved from a higher admission count; those outcomes require separate evidence.

    Where Workhint fits

    Workhint’s Workflow Engine describes defining request, approval, assignment and work-step movement, including required details, files or notes before progression. That is relevant to collecting an admission packet and routing scope review.

    The Workflow Automation Software solution describes configured forms, stages, owners, tasks and role-based coordination. Those capabilities provide a basis for discussing evidence and acceptance queues.

    These pages do not establish native incident correlation, automatic discovery of every relevant record, a CMDB, synchronized updates to an external service desk or immutable audit storage. Validate those requirements separately. Where a critical effect remains manual, retain an accountable verification step instead of assuming a notification changed the source system.

    Build This in Workhint

    Use this as a configuration brief to validate, not a claim that every proposed control is native:

    Design a problem-investigation admission system with coordinator, incident owner, analyst, service owner, evidence supplier and backup roles. Restrict evidence visibility and transition authority by duty. Intake the service, incident references, environment, observations, competing explanations and existing-investigation leads. Maintain a per-incident membership assessment with supporting and conflicting evidence.

    Use Signal captured, Evidence pending, Scope review, Acceptance pending, Admitted/linked, Returned/not admitted and Disputed/split review. Require accepted scope, qualified ownership, accessible evidence and verified relationship records before completion. Keep incident recovery and closure outside the admission decision.

    Create evidence, review, acceptance and link-verification tasks with event notifications. Give missing data, rejected scope, unavailable owners, duplicate investigations, conflicting evidence, failed handoffs and vendor waits an owned blocking state, escalation and resume condition. Provide views for aged signals, provisional membership, disputed links and unaccepted transfers. Measure demand, queue age, capacity, packet quality and verified handoff outcomes. Retain attributable decisions, earlier membership assessments, evidence references and audit history; validate retention, external updates and any unsupported controls before implementation.

    FAQ

    Must several incidents occur before opening an investigation?

    Not necessarily. Local policy may admit a single significant incident or a proactive service concern. Define the trigger and evidence rather than treating a universal count as a requirement.

    Does linking incidents prove they have the same cause?

    No. Record whether the relationship is provisional, supported, disputed or excluded. The link is a managed investigation relationship; causal conclusions require evidence.

    Should a linked incident be closed automatically?

    No. Its response owner must apply the incident’s own recovery and closure conditions. Admission to an investigation does not establish restoration.

    What finishes an admission case?

    An accepted investigative scope or accepted return destination, verified record relationships, retained reasoning and a named next action. Sending an assignment message alone is insufficient.

    Comments

    Leave a Reply

    Your email address will not be published. Required fields are marked *


    The reCAPTCHA verification period has expired. Please reload the page.