Skip to content

Monthly Checks

Applies to: All subscriptions

Purpose

Perform the checks that need a full cycle of data to be meaningful, and the hygiene tasks whose absence causes slow, hard-to-attribute decay.

Audience

The operations owner, with IT and finance.

Prerequisites

Steps

Schedule this shortly after the month closes, so the month's data is complete but any finding can still be acted on before filing.

1. Reconciliation

Verify Documents in the ERP against documents in Complifly, per registration, for the month
Expect Agreement, with every difference explained
Warning signs Documents in the ERP with no Complifly record; documents registered that the ERP does not show as sent; unexplained differences
Act Investigate every difference. A missing document is a compliance gap; an unexpected one is an integration fault

This is the most important monthly check. It is the only one that proves the two systems actually agree, and it is what catches an integration that has been silently dropping a subset of documents.

2. Compliance completeness

Verify Every document requiring registration has one; every movement had a valid bill
Expect Complete
Warning signs Documents still unregistered from the closed month; bills that expired mid-transit
Act Register what can still be registered. Anything past its window goes to the tax lead — it cannot be fixed here

3. Access review

Verify The user list against current staff; roles against current duties; registration access against need
Expect Every account belongs to a current employee with a current need
Warning signs Accounts for departed staff; accounts unused for 90 days; more administrators than necessary; fewer than two
Act Disable departed users — never delete, which would break attribution of their past actions. Trim excess access. Ensure at least two administrators

4. Audit trail review

Verify Administrative and control-relevant actions for the month
Expect Every one attributable and expected
Warning signs Configuration changes nobody recognises; segregation of duties relaxed; feature flags changed without a record; cross-tenant access used
Act Reconcile against the change log. Any control-relevant change without a record needs an explanation

5. Restore rehearsal

Verify A backup restores into a scratch environment and the application opens against it
Expect Restore completes; data present and correctly scoped
Warning signs Restore fails; takes longer than the recovery objective; data present but invisible to users
Act Fix before it is needed. Data present but invisible after a restore is a scope-identifier mismatch — reconcile the identifiers, never disable the isolation policies

A backup that has never been restored is an assumption. Quarterly is the minimum; monthly is better.

6. Database maintenance

Verify Index maintenance, statistics updates and integrity checks all ran
Expect All completing successfully
Warning signs Jobs failing; fragmentation rising; integrity check errors
Act Integrity errors are urgent — assess how far back they extend. Fragmentation and stale statistics are the usual cause of gradual slowdown

7. Retention and housekeeping

Verify Retention jobs ran; storage growth matches forecast; caches and temporary files are bounded
Expect Growth tracking the forecast
Warning signs Growth above forecast; retention jobs failing; dead-lettered events approaching their retention limit
Act Re-forecast if growth diverges. Investigate dead-lettered events before they are purged. See Housekeeping

8. Master data hygiene

Verify Counterparty registration statuses, duplicate parties, unused items, rate effective dates
Expect Clean
Warning signs Cancelled counterparty registrations still active in the master; duplicate parties; rates with no future validity
Act Cancelled counterparties cause registration failures at the worst moment. Verification results age, so re-check before high-value supplies

9. Documentation and knowledge

Verify Runbooks match reality; the deputy can perform the daily routine; contact lists are current
Expect Current
Warning signs Procedures nobody follows; a deputy who has never performed the routine; alerts addressed to departed staff
Act Update. Have the deputy actually run the daily routine this month

Validation

Check Pass condition
Reconciliation complete Every difference explained
Compliance complete Nothing unregistered from the closed month, other than statutory exceptions
Access reviewed Departed users disabled; at least two administrators
Audit reviewed Every control-relevant change accounted for
Restore rehearsed Completed within the recovery objective
Maintenance ran Index, statistics and integrity all successful
Growth on forecast Or the forecast revised
The deputy is capable Demonstrated this month

Troubleshooting

Symptom Cause Action
Reconciliation differences every month An integration dropping a subset Identify the pattern — a document type, a registration, a time of day
Documents unregistered past the window Ageing not checked daily Move it to the daily routine. It cannot be recovered afterwards
Departed users still active No link between the leaver process and Complifly Add Complifly to the leaver checklist
Restore takes longer than the objective Objective set without testing Either improve the process or revise the objective honestly
Restored data invisible to users Scope identifiers differ between environments Reconcile the identifiers. Never disable isolation
Maintenance jobs failing unnoticed Not monitored Alert on job failure
Storage above forecast Volume or retention differs from the plan Re-forecast and adjust retention deliberately
Registration failures for a long-standing customer Their registration was cancelled after you last checked Re-verify counterparties periodically
The deputy cannot run the routine Never practised Practise monthly, not at the point of need