Industrial network monitoring and fault analysis
For intermittent communication dropouts, we compare PLC and HMI logs, switch counters and downtime records on a single timeline. For recurring problems, we define which data is monitored and at what interval.
Port error countersPassive captureWiresharkAlertsRoot cause report
What problem does this service solve?
Downtime blamed on “the network” but never proven
Because nobody measures at the moment of failure, the argument goes round in circles between automation, IT and the machine manufacturer.
Error counters quietly increasing
CRC errors, frames discarded on the interface and the link going down and up (link flap) are separate findings and can accumulate in the counters for weeks. Without monitoring, these counters are only looked at after downtime.
Broadcast storms and loops
A single wrongly connected cable, or even an unmanaged switch added later, can create a loop; broadcast traffic rises and communication may be cut off.
The monitoring system not seeing OT
Enterprise monitoring sees the servers; field switches and PROFINET devices are often left out of monitoring.
What we cover, and within what limits
Monitoring means continuously recording the health of active network devices, port statistics, topology changes and communication quality; the aim is to raise an alert when a defined threshold is exceeded. How quickly a problem is noticed depends on the sampling interval and the defined detection condition. Fault analysis, on the other hand, isolates the cause of a specific outage by measurement: the physical layer (cable, connector, grounding), configuration (duplicate addresses, wrong VLAN, speed/duplex mismatch), network load (broadcast and multicast traffic) or the application side. We do traffic capture with port mirroring or a TAP; because port mirroring requires a configuration change on the switch and a TAP requires briefly interrupting the link, we plan the installation. Monitoring traffic can also load the control network, so we determine polling intervals and monitoring points by design. If the source of a dropout on the line side extends to the campus backbone, examine the shared link with enterprise network performance analysis.
Technical scope
Symptoms and timeline
Alarm logs, downtime times and affected devices are matched on a single event timeline.
Network device and port analysis
Port error counters, link changes, MAC/ARP tables and redundancy event logs.
Passive traffic analysis
Capture from the right point; broadcast/multicast ratio, retransmissions, PROFINET alarm and error frames.
Physical layer check
Cable route, connectors, shielding and grounding. If needed, cable testing with a certification tester (depending on scope).
Continuous monitoring setup
SNMP/LLDP-based monitoring, thresholds and alerts that reach the right person; integrated into the existing monitoring system if required.
Root cause report
Findings, evidence, recommended fixes and what to look at if the fault recurs.
Common symptoms and where to look first
| Symptom | Likely cause | First check |
|---|---|---|
| Intermittent dropout at a single station | Connector, cable, grounding, EMI, speed/duplex mismatch | CRC and discard counters of the relevant port, speed/duplex setting and cable route |
| Dropout at several stations at the same time | Backbone, redundancy event, switch power supply | Ring/redundancy event logs and switch power status |
| Network-wide slowdown or lock-up | Loop, broadcast storm, wrongly connected device | Broadcast/multicast ratio and STP/topology changes |
| No communication after a device replacement | Device name/IP, GSDML or VLAN mismatch | Device name assignment and port VLAN configuration |
| HMI screens updating slowly | Application load or network latency | Number of HMI–PLC connections and latency along the network path |
Prerequisites: what we need to get started
We list any gaps together in the first meeting and complete them during the survey.
- Alarm/downtime logs showing fault times
- Read-only access to managed switches (SNMP, web or console)
- Permission for port mirroring, or a TAP point, for passive capture
- A network diagram (if there is none, producing one is added to the scope)
- A contact who can be reached during the shift in which the problem is observed
Deliverables
Event timeline
Measurement data and traffic analysis findings
Evidence-based root cause report
Prioritised remediation plan
Monitoring dashboards, thresholds and alert list
A checklist describing the steps to follow at the moment of a fault
Representative scenario: a line that locks up several times a week
Several times a week, the line loses communication completely for a few minutes. If the switch logs show a topology change and a sudden spike in broadcast traffic coinciding with the lock-ups, a loop formed through an unmanaged switch added to the line later is the first possibility to verify. If the loop is confirmed, the connection is corrected, loop protection is enabled and a threshold alert is defined for the broadcast rate.
This is a representative analysis flow; it is not a real customer case.
Service summary
PROFINET and PROFIBUS Communication
We investigate intermittent dropouts between PLCs, remote I/O and drives using device diagnostic logs, port counters and the field topology.
Enterprise network monitoring (LAN/WAN) ↗
Scope limit
The access needed for the analysis must be provided, and the production lead must be on site during measurements. Fixing errors in machine and PLC application software is out of scope; we share our findings with the person responsible.
How do we proceed?
No step that could affect production is carried out until the maintenance window, configuration backup and rollback plan are clear.
Timeline
Symptoms and affected devices are matched.
Measurement
Port data and passive captures are collected.
Isolation
Physical, configuration, load and application causes are separated.
Remediation
Planned intervention and re-measurement are carried out.
Monitoring
Thresholds and alerts are set up; the sampling interval and detection condition are documented.
Questions about Industrial Network Monitoring and Fault Analysis
Does measuring affect production?
No packets are sent into the control traffic during capture. Installation, however, requires changes: port mirroring needs configuration on the switch and can add load to it, and inserting a TAP briefly interrupts the link. We plan both with the plant team.
Can we use our existing monitoring system?
Often, yes. OT switches can be added to enterprise monitoring systems that support SNMP and LLDP; we adjust polling intervals and access rules with the load on the OT network in mind.
How do you catch a fault that occurs rarely?
We leave traffic capture and counter monitoring running for a set period; when the fault recurs, we have the data from that moment. The period depends on how often the fault occurs.
What happens if the root cause is not found?
We report the causes ruled out by measurement and the remaining possibilities, and recommend the next measurement step. We do not guarantee results; we document which cause was eliminated by which measurement.
Work that can be planned together
PROFINET and PROFIBUS Communication
We investigate intermittent dropouts between PLCs, remote I/O and drives using device diagnostic logs, port counters and the field topology.
Industrial Ethernet and Factory Networks
We map switches, cables and connected devices from machine cabinets to the plant backbone.
Factory OT Infrastructure Discovery
Before any investment, we record on site which PLCs, switches, protocols and data flows are in use.
Knowledge Center
Technical guides for this service
Let's plan a technical survey for industrial network monitoring and fault analysis
Briefly describe the line, the PLC and network equipment you use and the problem you want to solve; we will respond with a written proposal covering scope, assumptions and deliverables.