Cisco & Meraki specialists. Across the UK.
+44 203 196 6566 · option 1sales@proxar.co.uk

Connect. Secure. Manage.

BGP Troubleshooting

Investigate Cisco BGP session failures, missing prefixes and unexpected path selection using neighbour, policy and forwarding evidence.

Cisco Partner since 2009 Projects & ongoing support

Illustrative Cisco and Meraki equipment in organised data-centre racks
Your network. Properly considered.Design · Implementation · Support
1,000+

Devices supported
across the UK

20+

Client locations
supported

Nearly 8 Tbps

Traffic handled
at peak time

3 min

Average
response time

Average response is not a resolution time or a contractual guarantee. Support scope and response commitments are agreed for each engagement.

BGP troubleshooting starts by separating a failed session, a missing route and a forwarding problem. They can produce similar user symptoms, but they require different evidence and different corrective actions. MCI Network Services investigates Cisco BGP incidents for internet edges and private interconnects, working with customer teams and providers to identify the failure without making broad, unreviewed changes to a live routing policy.

01

Cisco-focused engineering

Design around the platform, traffic and failure behaviour you actually operate.

02

Change with evidence

Use pre-checks, acceptance criteria and a documented route back.

03

Operational handover

Keep the configuration, diagrams and responsibilities understandable after delivery.

Define the symptom and its boundaries

We establish what stopped working, when it was first observed and which destinations or applications are affected. A provider alarm, a route withdrawal and an application error may occur at different times. We record those distinctions rather than assuming the first alert identifies the beginning of the fault.

Useful starting information includes a redacted topology, affected prefixes, peer addresses and recent changes. We also ask whether one direction of traffic still works and whether a backup path is carrying service. This helps distinguish a complete connection loss from a policy change affecting selected networks. Sensitive configuration and captures should use an agreed transfer route, not an unrestricted public form.

Illustrative Cisco and Meraki industrial network equipment in a plant environment
01Network engineering for operational and industrial environments.
Cisco PartnerCisco Partner
Cisco MerakiCisco Meraki
ISO 9001Quality management
ISO 27001Information security
Cyber EssentialsCyber Essentials
Cyber Essentials PlusCyber Essentials Plus

MCI Network Services is operated by Proxar Ltd. Certifications apply to the certified legal entity and the scope of its certificates. View accreditation information

Check the session before changing the routing policy

For a session that is not established, we examine underlying reachability, interface state, addressing, the TCP connection and peer configuration. Autonomous-system mismatches, authentication issues, access controls and multihop prerequisites can all prevent a session from forming. The investigation follows the observed state instead of guessing which command might restore it.

We compare evidence from both ends where possible. A local timeout can have several causes, and the remote team may see a more specific rejection. Provider coordination is therefore part of the service, but it does not transfer the carrier’s repair obligations to MCI. We make the evidence clear enough for the responsible team to act.

Illustrative Cisco and Meraki data-centre network infrastructure
02Network infrastructure designed around the workloads it carries.

A useful starting point

Not sure where to start?

Bring the requirement. We will help you define the engineering work.

Talk to a specialist

Trace a missing prefix through the policy

An established neighbour does not mean that a particular prefix has been accepted. We check whether the route is being originated, advertised by the peer, received locally, permitted by policy and selected for installation. Prefix filters, maximum-prefix events, address-family configuration and validation policy are reviewed at the relevant stage.

We avoid removing all filters to test a theory. A controlled policy examination or narrowly scoped test is safer and produces better evidence. Where a provider rejects a legitimate advertisement, we compare the intended origin and registration information with its requirements. The correction may be administrative rather than a router configuration change.

Illustrative Cisco and Meraki network rack in a modern office
03The office network: switching, uplinks and structured cabling.

Separate the BGP table from forwarding

A route can exist in BGP without being the active route used for traffic. Another source may win, a next hop may be unreachable or a more-specific prefix may direct traffic elsewhere. We inspect the installed routing and forwarding state, not only the list of received BGP paths.

The return direction matters too. Stateful firewalls, NAT and private partner policies can make an asymmetric route fail even when every router has a destination entry. We trace representative traffic through those boundaries and distinguish network reachability from DNS, entitlement or application problems. In market-data environments, a healthy BGP session does not establish that multicast delivery or feed recovery is healthy.

Illustrative enterprise Cisco network rack in a modern office environment
04Enterprise switching and routing, built to be operated.

Investigate intermittent changes

Flapping sessions and unstable paths need a time-based investigation. We correlate routing events with interface errors, circuit alarms, control-plane load and configuration activity. BFD events, where deployed, are examined alongside BGP rather than treated as proof of the underlying cause on their own.

We also consider whether an automatic preference change is behaving as designed but exposing an application dependency. For example, a backup path may use a different translated source address. The resulting failure is not necessarily a BGP defect. The investigation should explain the relationship between the route event and the business symptom.

Illustrative Cisco and Meraki fibre network rack with structured cabling
05Fibre, switching and the physical path behind the service.

Restore service without losing the evidence

An urgent incident may require a temporary workaround before the full cause is established. We agree the scope, risk and rollback with the customer, retain the pre-change evidence and record why the action was taken. A workaround that restores connectivity is labelled as such until its long-term suitability is reviewed.

Broad session resets can interrupt unrelated traffic and erase useful state. We use them only when justified by the evidence and change authority. Supported refresh or targeted actions may be less disruptive, but their effect is checked for the actual platform. The aim is to restore useful service while preserving a defensible explanation of what happened.

Illustrative Cisco and Meraki equipment in organised data-centre racks
06Data-centre connectivity with clear physical and logical boundaries.

Finish with findings the business can use

The investigation report separates confirmed facts, plausible contributing factors and unresolved questions. It identifies any provider action, configuration correction or monitoring improvement still required. Where resilience was lost without a total outage, that is recorded as a finding rather than dismissed because users eventually reconnected.

MCI can provide a focused incident engagement or a broader BGP routing review after stabilisation. Service hours, response commitments and access arrangements are agreed for the engagement. The outcome should be a clearer network and a practical prevention plan, not simply a closed ticket saying that the neighbour is up again.

Illustrative Cisco and Meraki industrial network equipment in a plant environment
07Network engineering for operational and industrial environments.

Talk to a specialist

What does your network need next?

Tell us what you are planning, changing or trying to improve. We will help define a practical next step and the engineering scope.

+44 203 196 6566

Choose option 1

sales@proxar.co.uk

UK-wide service.
Most of our engineers are based in Kent, with a focus on London, Sussex, Essex and Berkshire.

Existing customers should use their agreed support and escalation channels. This form is for service enquiries, not a monitored emergency ticket.

Tell us what you need.

A short outline is enough. Please do not include passwords or sensitive network details.

Thank you. Your enquiry has been submitted.

We’ll use these details to respond to your enquiry. Please leave out passwords and sensitive network information. Read our privacy notice.