IT systems · Practical guide

Validating elproMONITOR: from a recorded reading to a timely response

A complete elproMONITOR trace cannot show whether a responsible person received an alarm in time. Recording, alarm delivery, acknowledgement and response can diverge during maintenance or a connection failure. The timing example distinguishes documented 3.3.x and cloud features and identifies evidence for QA, operations and IT.

Table of contents

At a glance

How do you validate elproMONITOR for your monitoring process?

Identify sensors, modules, software version, deployment model and notification routes. Check whether relevant conditions are detected, reach the correct recipients and receive a documented response. Include disabled sensors, communication failures and the configured acknowledgement process. Use relevant supplier evidence. Actual requirements, results and the applicable GMP or GDP scope determine whether your monitoring process is suitable.

  • elproMONITOR 3.3.x and Cloud have separate documentation areas; capture the actual operating status.
  • Logging may continue, even if warning alarms are disabled.
  • Retrospective data transfer indicates a lack of timely notification during a system failure.
  • Events, clearance, and technical handling are separate pieces of evidence.

1. Define the monitored process and its control strategy linkage

elproMONITOR is ELPRO's software solution for room and equipment monitoring. The description covers connected measurement modules, central display, alarm routing, and user management. Therefore, the system boundary extends beyond the server: it includes the sensors used, communication modules, networks, notification services, and the controlled handling of alerts. The same software can be applied to a storage area and a production device, each with distinct operational requirements.

Under EU GMP, Annex 11 is the applicable guideline for computerised systems in the manufacture of human medicinal products. In the human medicinal product wholesale sector, GDP guidelines must be independently considered. Section 3.3.1 of GDP outlines appropriate validation or verification evidence prior to use. Therefore, GDP is not merely synonymous with EU GMP, even though specific evidence may be meaningfully aligned.

From an operational perspective: What deviation must be detected, who requires the information, and what action should remain feasible in a timely manner? Only from this basis can appropriate limits, response times, and error scenarios be established. The product name or a general compliance statement does not address these questions.

Sources: [1] [7] [9] [6]

Further reading:

2. 3.3.x, distinguish cloud and actual configuration

The official documentation includes separate sections for elproMONITOR 3.3.x on-premise and cloud environments, as well as for older version lines. The presence of a handbook title does not constitute confirmation of an active build. Capture application, modules, firmware, identity management, and operating mode. Also document the specific document version used for a requirement or test.

ELPRO specifies comparable core functions for both cloud and on-premises environments, while also including tailored features, user interfaces, and workflows. Therefore, a user instruction from the 3.3.x manual cannot be assumed to represent identical cloud functionality without proper validation. Existing hardware must be assessed for compatibility during transition to new systems. A functional sensor connection alone does not confirm the integrity of all alarm and data pathways.

Assign existing supplier documentation to the actually utilized functions. Appropriate installation and functional tests can cover a significant portion of the requirements. Additional evidence may include alarm profiles, recipients, rights, location bindings, and failover procedures. The use of existing documentation requires a rationale-based alignment review, not a blanket reconstruction of all documents.

Sources: [2] [5] [1] [7]

3. Evaluate stored values and active alarmings separately

The cloud handbook section on deactivation describes a particularly important distinction: measurement data continues to be recorded, while warning limits are suppressed and the data is not included in statistical evaluations. Such a state may be planned for maintenance activities. However, it must have a clear purpose, appropriate authorizations, and a traceable closure procedure.

The following internal work matrix assists in interpreting seemingly contradictory alerts in a technically accurate manner. Always verify the actual system configuration, rather than assuming uninterrupted operation based solely on an existing curve or an empty alarm view.

Own work matrix: What the four states actually intend to convey
StateImportant differenceTo demonstrate the project
Maintenance sensor disabled Recording is ongoing, alarm notification is limited Authority, Comment, Period, Alternative Measure and Reactivation.
Limit value again upheld Event completed, processing may be pending The relevant notification and intended clearance remain appropriately traceable.
Connection interrupted Local data may be available; remote information is missing. Actual available local alarm notification pathways and predefined failure response process.
Data synchronized later Historical completeness improves; prior reactions are not retrospectively captured Event time, transmission, and reaction are evaluated separately.

Sources: [4] [3] [5] [7]

4. Verify alarm profiles, recipients, and time reference explicitly

Start with the actual alarm profiles. For each relevant state, the professional boundary, if applicable, delay, recipient, and expected response should be clearly defined. A temperature event is a distinct case from a lost connection or a battery warning. The 3.3.x Handbook distinguishes between these respective alarm types. Their priority and handling are determined by the intended use.

Verify an end-to-end pathway to the reachable recipient. A generated email does not constitute evidence that the on-call service receives it and can assign it. Consider real-world distribution lists, shift changes, inactive users, and failed transmission services. Justify which of these scenarios are tested based on their potential impact.

Use a common time base when comparing time. The 3.3.x manual describes ads according to the selected user time zone. For your detection, event, notification and reaction should therefore be clearly comparable. Document deviations between scheduled and observed sequence instead of just checking a successful message display.

Sources: [3] [1] [7] [9]

Further reading:

5. Example: Data available, report submitted late

The following example does not represent a completed elproMONITOR test. All times are in the same time zone. A deviation period begins at 12:00 and ends at 12:07. For this test process, the responsible recipient must be informed no later than five minutes after the start. This deadline is not a drug-specific specification and does not constitute a general regulatory requirement.

Additionally, an interrupted remote connection from 12:01 to 12:10 is assumed. The example includes a remote message at 12:11 and an initial response at 12:13. It deliberately describes a case to be evaluated, not a specified product latency. A local alternative pathway would need to be separately demonstrated.

Example: fully calculated workflow with own acceptance criterion
EventIndependent calculationEvaluation and Demonstration
Start at 12:00; information objective to be achieved within 5 minutes 12:00 + 5 minutes = 12:05 The recipient must be informed of this test process by 12:05.
Connection lost from 12:01 to 12:10 12:10 − 12:01 = 9 minutes Local data entry and potential backup alarm must be separately justified.
Limit violation ends 12:07 12:07 − 12:00 = 7 minutes Normal values thereafter do not invalidate the previous finding.
Assumption: Remote message received by recipient at 12:11 11 minutes after start; 6 minutes after the scheduled time The intended purpose is not met, if no appropriate alternative method was available in a timely and effective manner.
Assumption: First reaction occurs at 12:13 2 minutes after notification; 13 minutes after event onset Document reaction time separately from transport and recognition time.
A complete calibration curve appears subsequently. The time balance of delayed notifications remains unchanged Data completeness cannot retrospectively replace previously missing information.

Sources: [5] [7] [9]

6. Receipt and technical processing not to be confused

In the 3.3.x Handbook, the setting of MustConfirmAlarms determines how unacknowledged alarms are handled in the respective view. The documented acknowledgment includes timestamp and user information. Variants with and without electronic signature are described. For the signature variant, the required number of acknowledging users is configurable.

Do not derive a general two-person control requirement from this. Define the role and significance of a confirmation within your process: reception confirmed, root cause investigated, or corrective action completed—these represent different interpretations. A technical confirmation should not be implicitly treated as comprehensive professional reassurance. Furthermore, a return to normal measurement value alone does not constitute sufficient justification for releasing affected batches.

Review the relevant views and filters under different states. An alert may disappear from a work view without being deleted. For future evaluations, the necessary linkage between event, user action, comment, and corrective measure must remain identifiable. The actual signature requirements are to be determined at the specific procedure level.

Sources: [3] [7] [9]

7. Calibration, evaluation, and documentation to be maintained together

A calibration qualification addresses questions regarding the metrological suitability of the relevant device. It does not demonstrate that an alarm profile is correctly assigned or that the designated recipient is reached. The cloud handbook section describes a dedicated calibration state, as well as the deactivation and activation of monitoring. Verify how these actions are controlled within your intended operating procedure.

For stationary GDP monitoring, Section 3.3 specifies calibration intervals based on risk and reliability. A universal annual schedule for all measurement devices should not be derived. Specific device type, usage, and applicable requirements must be considered. Section 3.3 also expects appropriate alarm systems with relevant alarm thresholds and regular functional checks. Therefore, in GDP operations, plan and document recurring alarm tests, in addition to those conducted during installation or changes. Calibration and alarm testing address different questions.

Determine additionally which records shall support future evaluation: measurement data, relevant profiles, deactivated time periods, alerts, and processing information. Within the GMP scope, Annex 11 and Chapter 4 assist in distinguishing between change documentation and recordkeeping. A single PDF report does not automatically demonstrate that all required relationships are preserved.

Sources: [4] [9] [7] [8]

8. Cloud operation and evaluation of impact of changes

The Cloud FAQ explicitly states that local processing and subsequent synchronization during network interruptions depend specifically on the configuration. Therefore, clarify the actual hardware setup, buffering capabilities, local alarm notification pathways, and the process for returning to normal operation. General statements regarding cloud availability do not address whether your specific alert is delivered in a timely manner during a network outage.

Separate shielding, recovery, and professional archiving. Required data shall remain usable throughout the relevant time period; during recovery, their relationship to configuration and processing must also be maintained. Define responsibilities and appropriate evidence with relevant service providers. Public hosting advertising does not constitute a specific service and recovery agreement.

Assess changes based on their impact: a new alarm profile, different recipients, updated identity management, hardware replacement, or cloud migration may require different testing approaches. Appropriate existing documentation may be reused. A full revalidation at every update is no more justified than the assumption that a supplier update automatically covers your entire process.

Sources: [5] [1] [7] [9]

9. What preparation determines the effort for SMEs

For a robust preparation, you need a list of measurement points and components, alarm profiles, recipient pathways, and existing documentation. Designate subject matter experts responsible for limits and actions, as well as point of contact for IT and on-call services. A manageable test case should illustrate the key operational decision pathway and not merely reflect the most attractive software interface.

Effort arises particularly from different locations, multiple alarm services, outdated hardware, individual user rights, and unclear failure response procedures. While many similar sensors can be more readily evaluated than a small number of measurement points employing entirely different reaction pathways, project scope based solely on sensor count often proves insufficient; specific time or cost commitments require a more detailed inventory and assessment.

Ask the supplier specifically for suitable version documents and tests. Then determine which customer-specific transitions will be additionally tested. Department and QA explain the use and its risks, IT the technical dependencies. Tasks and decision-making powers are regulated according to the organization, not taken over from a rigid general role list.

Sources: [1] [7] [9]

10. Checklist for demonstrable ongoing monitoring operation

Use the following internal guidance document for your usage decision and for future amendments. A point is only considered resolved when the relevant information or a evaluated justification is available. Open findings require a responsible party and a decision on their impact on the intended operation.

  • Operating mode, software version, modules, and communication pathways are clearly documented.
  • GMP, GDP, or any other specific intended use and relevant decision-making activities are clearly defined.
  • Alarm profiles, time reference, recipients, and required actions are professionally defined.
  • Deactivation, calibration, and reactivation are controlled and traceable.
  • Interruption and failure of a notification pathway are appropriately evaluated.
  • Events, closures, and professional handling have clear significance.
  • Required data, modifications, and reports remain accessible for the intended use.
  • Supplier records, restoration, and change evaluation align with the actual configuration.
  • Periodic functional tests of GDP alarmings are defined and documented.

Sources: [7] [9] [8]

Project experience and collaboration

Follow the alarm through to the operational response

Monitoring is effective only if the responsible person can act in time. With Operations and IT, we trace a deviation from detection to notification, acknowledgement and substantive handling. A recipient list or alarm profile may need a targeted change and retest. Communication failures or unclear response paths require an agreed operating procedure. We base requirements on the storage or process conditions being monitored and the elproMONITOR components actually installed.

Frequent Questions about elproMONITOR and Alarm Paths

Has elproMONITOR been fully validated as a complete process?

ELPRO provides software and qualification support. This does not constitute a general confirmation of your alarm profiles, recipients, failure pathways, and operational responses. Utilize existing evidence based on their actual scope and supplement with relevant customer-specific transition points. For GMP and GDP, the specific application is decisive; general manufacturer marketing does not substitute for this alignment.

Sources: [1] [7] [9]

Are elproMONITOR 3.3.x and Cloud functionally identical?

This should not be assumed. The documentation separates both areas, and the Cloud-FAQ describes comparable core functions with adapted features and workflows. Assess the actual operating setup and the documents currently in use. When switching, particularly verify hardware compatibility, rights, data transfer, and alarm routing. A similar screen display does not constitute a complete proof of equivalence.

Sources: [2] [5]

Why may measurement values be available even though no limit alarm has been triggered?

The Cloud Handbook section states that during deactivation, data recording continues, but limit alarm notifications cease and values are not included in the statistics. Therefore, the actual sensor status and the relevant time period must be verified. Whether additional root causes exist must be investigated specifically. A measurement curve alone neither confirms active alarm triggering nor demonstrates timely operational notification.

Sources: [4]

Does subsequent synchronization indicate a successful failure test?

It can prove that certain data became available later. This is not enough for a time-critical alarm function alone. In addition, check which information was available locally and remotely during the failure and when a responsible receiver was reached. The cloud FAQ describes the behavior depending on the configuration. A general claim to timely remote delivery cannot be derived from this.

Sources: [5] [7] [9]

What is the impact of MustConfirmAlarms?

In the 3.3.x Handbook, the setting affects the handling of unresolved notifications within the corresponding view. For example, for a no longer active notification, it is essential to determine whether confirmation is required. Verify the actual filter and the configured rule. The disappearance from a work view must not be automatically interpreted as data deletion or a fully completed root cause assessment.

Sources: [3]

Is every approval considered an electronic signature?

No. The 3.3.x Handbook explicitly distinguishes between clearance with and without electronic signature. For the signature variant, the number of required users can be configured. Define your mode and the meaning of the action. Whether a signature is required and appropriate for a specific procedure depends on its regulated context, not merely on the presence of a confirmation button.

Sources: [3] [7]

Must all stationary sensors be calibrated annually?

Such a general statement cannot be derived from the GDP guideline for storage monitoring referenced herein. Section 3.3 specifies defined intervals based on risk and reliability. Determine the appropriate interval based on the device, usage, and applicable requirements. Other activities, such as transport monitoring, require separate consideration. Calibration does not substitute for verification of the alarm pathway.

Sources: [9] [7]

Which recipients should be included in the alarm test?

Include the intended recipients and representatives for your process. Crucially, determine who is actually capable of detecting and addressing a report during the relevant operating hours. Assess appropriate exception scenarios, such as invalid distribution lists or failure of a used service. The selection constitutes a risk-based evaluation plan; there is no general obligation to individually test every conceivable individual.

Sources: [1] [7] [9]

Does a transition to the cloud always require a complete revalidation?

The transition requires an assessment of its actual impact. Existing, relevant evidence can be reused; however, changes to data pathways, rights, operational responsibilities, or functions require appropriate supplements. ELPRO outlines migration support and highlights potential hardware and functional differences. A complete replication of all tests or automatic transferability of all legacy data should not be assumed.

Sources: [5] [7] [9]

What should we bring to the first project meeting?

Useful are measurement point lists, component status displays, alarm profiles, receiver pathways, existing test samples, and known faults. Supplement with the technically required response and your maintenance or connection loss procedure. This enables targeted identification of critical transition points and missing information. Effort and sequence can then be planned based on your actual configuration, without blanket commitments regarding run time or results.

Sources: [1] [7] [9]

Sources and review date

Source date: 6 October 2026. Manufacturer documentation from separate documentation domains for 3.3.x and Cloud; no current customer build number confirmed. Cloud behavior during connection loss is configuration-dependent. EU GMP Annex 11 and Chapter 4, both issued in January 2011, are applied only within the relevant human medicinal product GMP scope; Human Medicinal Product GDP 2013/C 343/01 is separately categorized. Time values and target matrices are subject to independent assumptions.

  1. ELPRO: elproMONITOR – Room & Equipment Monitoring
  2. ELPRO Online Manuals: elproMONITOR (EN)
  3. ELPRO Online Manuals: elproMONITOR 3.3.x / Deviations (EN)
  4. ELPRO Online Manuals: Cloud – Monitor/Sensors (EN)
  5. ELPRO: elproMONITOR Cloud – published webinar FAQ
  6. European Commission: EudraLex Volume 4
  7. EU GMP Annex 11: Computerised Systems
  8. EU GMP Chapter 4: Documentation
  9. European Commission: GDP Guidelines for Human Medicinal Products

Sources checked: .

Daniel Herrmann, Daniel Herrmann Consulting

Daniel Herrmann Consulting

Your alarm chain deliberately verify

We support you in aligning professional response requirements, existing supplier documentation, and the actual technical pathways.

  • Computer system validation →

    Bring your GxP systems into validated operation, from laboratory software, MES and eQMS to ERP. We define the scope, plan risk-based tests using GAMP 5 and hand the evidence to QA for release.

Send a project enquiry Customer testimonials →