ExpressRoute monitoring, metrics, and alerts

This article helps you understand ExpressRoute monitoring, metrics, and alerts using Azure Monitor. Azure Monitor is one stop shop for all metrics, alerting, diagnostic logs across all of Azure.

Note

Using Classic Metrics is not recommended.

ExpressRoute metrics

To view Metrics, go to the Azure Monitor page and select Metrics. To view ExpressRoute metrics, filter by Resource Type ExpressRoute circuits. To view Global Reach metrics, filter by Resource Type ExpressRoute circuits and select an ExpressRoute circuit resource that has Global Reach enabled. To view ExpressRoute Direct metrics, filter Resource Type by ExpressRoute Ports.

Once a metric is selected, the default aggregation will be applied. Optionally, you can apply splitting, which will show the metric with different dimensions.

Important

When viewing ExpressRoute metrics in the Azure portal, select a time granularity of 5 minutes or greater for best possible results.

Screenshot of time granularity options.

Aggregation Types:

Metrics explorer supports SUM, MAX, MIN, AVG and COUNT as aggregation types. You should use the recommended Aggregation type when reviewing the insights for each ExpressRoute metric.

  • Sum: The sum of all values captured during the aggregation interval.
  • Count: The number of measurements captured during the aggregation interval.
  • Average: The average of the metric values captured during the aggregation interval.
  • Min: The smallest value captured during the aggregation interval.
  • Max: The largest value captured during the aggregation interval.

ExpressRoute circuit

Metric Category Unit Aggregation Type Description Dimensions Exportable via Diagnostic Settings?
Arp Availability Availability Percent Average ARP Availability from MSEE towards all peers. Peering Type, Peer Yes
Bgp Availability Availability Percent Average BGP Availability from MSEE towards all peers. Peering Type, Peer Yes
BitsInPerSecond Traffic BitsPerSecond Average Bits ingressing Azure per second Peering Type No
BitsOutPerSecond Traffic BitsPerSecond Average Bits egressing Azure per second Peering Type No
DroppedInBitsPerSecond Traffic BitsPerSecond Average Ingress bits of data dropped per second Peering Type Yes
DroppedOutBitsPerSecond Traffic BitPerSecond Average Egress bits of data dropped per second Peering Type Yes
GlobalReachBitsInPerSecond Traffic BitsPerSecond Average Bits ingressing Azure per second PeeredCircuitSKey No
GlobalReachBitsOutPerSecond Traffic BitsPerSecond Average Bits egressing Azure per second PeeredCircuitSKey No

Note

Using GlobalGlobalReachBitsInPerSecond and GlobalGlobalReachBitsOutPerSecond will only be visible if at least one Global Reach connection is established.

ExpressRoute gateways

Metric Category Unit Aggregation Type Description Dimensions Exportable via Diagnostic Settings?
Bits received per second Performance BitsPerSecond Average Total bits received on ExpressRoute gateway per second roleInstance No
CPU utilization Performance Count Average CPU Utilization of the ExpressRoute Gateway roleInstance Yes
Packets per second Performance CountPerSecond Average Total Packets received on ExpressRoute Gateway per second roleInstance No
Count of routes advertised to peer Availability Count Maximum Count Of Routes Advertised To Peer by ExpressRouteGateway roleInstance Yes
Count of routes learned from peer Availability Count Maximum Count Of Routes Learned From Peer by ExpressRouteGateway roleInstance Yes
Frequency of routes changed Availability Count Total Frequency of Routes change in ExpressRoute Gateway roleInstance No
Number of VMs in virtual network Availability Count Maximum Number of VMs in the Virtual Network No Dimensions No

ExpressRoute Gateway connections

Metric Category Unit Aggregation Type Description Dimensions Exportable via Diagnostic Settings?
BitsInPerSecond Traffic BitsPerSecond Average Bits ingressing Azure per second through ExpressRoute gateway ConnectionName No
BitsOutPerSecond Traffic BitsPerSecond Average Bits egressing Azure per second through ExpressRoute gateway ConnectionName No
DroppedInBitsPerSecond Traffic BitsPerSecond Average Ingress bits of data dropped per second ConnectionName Yes
DroppedOutBitsPerSecond Traffic BitPerSecond Average Egress bits of data dropped per second ConnectionName Yes

ExpressRoute Direct

Metric Category Unit Aggregation Type Description Dimensions Exportable via Diagnostic Settings?
BitsInPerSecond Traffic BitsPerSecond Average Bits ingressing Azure per second Link No
BitsOutPerSecond Traffic BitsPerSecond Average Bits egressing Azure per second Link No
DroppedInBitsPerSecond Traffic BitsPerSecond Average Ingress bits of data dropped per second Link No
DroppedOutBitsPerSecond Traffic BitPerSecond Average Egress bits of data dropped per second Link No
AdminState Physical Connectivity Count Average Admin state of the port Link No
LineProtocol Physical Connectivity Count Average Line protocol status of the port Link No
RxLightLevel Physical Connectivity Count Average Rx Light level in dBm Link, Lane No
TxLightLevel Physical Connectivity Count Average Tx light level in dBm Link, Lane No

ExpressRoute Traffic Collector

Metric Category Unit Aggregation Type Description Dimensions Exportable via Diagnostic Settings?
CPU utilization Performance Count Average CPU Utilization of the ExpressRoute Traffic Collector roleInstance Yes
Memory Utilization Performance CountPerSecond Average Memory Utilization of the ExpressRoute Traffic Collector roleInstance Yes
Count of flow records processed Availability Count Maximum Count of number of flow records processed or ingested roleInstance, ExpressRoute Circuit Yes

Circuits metrics

Bits In and Out - Metrics across all peerings

Aggregation type: Avg

You can view metrics across all peerings on a given ExpressRoute circuit.

circuit metrics

Bits In and Out - Metrics per peering

Aggregation type: Avg

You can view metrics for private, public, and Microsoft peering in bits/second.

metrics per peering

BGP Availability - Split by Peer

Aggregation type: Avg

You can view near to real-time availability of BGP (Layer-3 connectivity) across peerings and peers (Primary and Secondary ExpressRoute routers). This dashboard shows the Primary BGP session status is up for private peering and the Second BGP session status is down for private peering.

BGP availability per peer

Note

During maintenance between the Microsoft edge and core network, BGP availability will appear down even if the BGP session between the customer edge and Microsoft edge remains up. For information about maintenance between the Microsoft edge and core network, make sure to have your maintenance alerts turned on and configured.

ARP Availability - Split by Peering

Aggregation type: Avg

You can view near to real-time availability of ARP (Layer-3 connectivity) across peerings and peers (Primary and Secondary ExpressRoute routers). This dashboard shows the Private Peering ARP session status is up across both peers, but down for Microsoft peering for both peers. The default aggregation (Average) was utilized across both peers.

ARP availability per peer

ExpressRoute Direct Metrics

Aggregation type: Avg

You can view the Admin state for each link of the ExpressRoute Direct port pair. The Admin state represents if the physical port is on or off. This state is required to pass traffic across the ExpressRoute Direct connection.

ER Direct admin state

Aggregation type: Avg

You can view the bits in per second across both links of the ExpressRoute Direct port pair. Monitor this dashboard to compare inbound bandwidth for both links.

ER Direct bits in per second

Aggregation type: Avg

You can also view the bits out per second across both links of the ExpressRoute Direct port pair. Monitor this dashboard to compare outbound bandwidth for both links.

ER Direct bits out per second

Aggregation type: Avg

You can view the line protocol across each link of the ExpressRoute Direct port pair. The Line Protocol indicates if the physical link is up and running over ExpressRoute Direct. Monitor this dashboard and set alerts to know when the physical connection has gone down.

ER Direct line protocol

Aggregation type: Avg

You can view the Rx light level (the light level that the ExpressRoute Direct port is receiving) for each port. Healthy Rx light levels generally fall within a range of -10 dBm to 0 dBm. Set alerts to be notified if the Rx light level falls outside of the healthy range.

ER Direct line Rx Light Level

Note

ExpressRoute Direct connectivity is hosted across different device platforms. Some ExpressRoute Direct connections will support a split view for Rx light levels by lane. However, this is not supported on all deployments.

Aggregation type: Avg

You can view the Tx light level (the light level that the ExpressRoute Direct port is transmitting) for each port. Healthy Tx light levels generally fall within a range of -10 dBm to 0 dBm. Set alerts to be notified if the Tx light level falls outside of the healthy range.

ER Direct line Tx Light Level

Note

ExpressRoute Direct connectivity is hosted across different device platforms. Some ExpressRoute Direct connections will support a split view for Tx light levels by lane. However, this is not supported on all deployments.

ExpressRoute Virtual Network Gateway Metrics

Aggregation type: Avg

When you deploy an ExpressRoute gateway, Azure manages the compute and functions of your gateway. There are six gateway metrics available to you to better understand the performance of your gateway:

  • Bits received per second
  • CPU Utilization
  • Packets per seconds
  • Count of routes advertised to peers
  • Count of routes learned from peers
  • Frequency of routes changed
  • Number of VMs in the virtual network

It's highly recommended you set alerts for each of these metrics so that you're aware of when your gateway could be seeing performance issues.

Bits received per second - Split by instance

Aggregation type: Avg

This metric captures inbound bandwidth utilization on the ExpressRoute virtual network gateway instances. Set an alert for how frequent the bandwidth utilization exceeds a certain threshold. If you need more bandwidth, increase the size of the ExpressRoute virtual network gateway.

Screenshot of inbound bit per second - split metrics.

CPU Utilization - Split by instance

Aggregation type: Avg

You can view the CPU utilization of each gateway instance. The CPU utilization may spike briefly during routine host maintenance but prolong high CPU utilization could indicate your gateway is reaching a performance bottleneck. Increasing the size of the ExpressRoute gateway may resolve this issue. Set an alert for how frequent the CPU utilization exceeds a certain threshold.

Screenshot of CPU utilization - split metrics.

Packets Per Second - Split by instance

Aggregation type: Avg

This metric captures the number of inbound packets traversing the ExpressRoute gateway. You should expect to see a consistent stream of data here if your gateway is receiving traffic from your on-premises network. Set an alert for when the number of packets per second drops below a threshold indicating that your gateway is no longer receiving traffic.

Screenshot of packets per second - split metrics.

Count of Routes Advertised to Peer - Split by instance

Aggregation type: Max

This metric shows the number of routes the ExpressRoute gateway is advertising to the circuit. The address spaces may include virtual networks that are connected using VNet peering and uses remote ExpressRoute gateway. You should expect the number of routes to remain consistent unless there are frequent changes to the virtual network address spaces. Set an alert for when the number of advertised routes drop below the threshold for the number of virtual network address spaces you're aware of.

Screenshot of count of routes advertised to peer.

Count of routes learned from peer - Split by instance

Aggregation type: Max

This metric shows the number of routes the ExpressRoute gateway is learning from peers connected to the ExpressRoute circuit. These routes can be either from another virtual network connected to the same circuit or learned from on-premises. Set an alert for when the number of learned routes drop below a certain threshold. This could indicate either the gateway is seeing a performance problem or remote peers are no longer advertising routes to the ExpressRoute circuit.

Screenshot of count of routes learned from peer.

Frequency of routes change - Split by instance

Aggregation type: Sum

This metric shows the frequency of routes being learned from or advertised to remote peers. You should first investigate your on-premises devices to understand why the network is changing so frequently. A high frequency in routes change could indicate a performance problem on the ExpressRoute gateway where scaling the gateway SKU up may resolve the problem. Set an alert for a frequency threshold to be aware of when your ExpressRoute gateway is seeing abnormal route changes.

Screenshot of frequency of routes changed metric.

Number of VMs in the virtual network

Aggregation type: Max

This metric shows the number of virtual machines that are using the ExpressRoute gateway. The number of virtual machines may include VMs from peered virtual networks that use the same ExpressRoute gateway. Set an alert for this metric if the number of VMs goes above a certain threshold that could affect the gateway performance.

Screenshot of number of virtual machines in the virtual network metric.

Note

To maintain reliability of the service, Microsoft often performs platform or OS maintenance on the gateway service. During this time, this metric may fluctuate and report inaccurately.

ExpressRoute gateway connections in bits/seconds

Aggregation type: Avg

This metric shows the bits per second for ingress and egress to Azure through the ExpressRoute gateway. You can split this metric further to see specific connections to the ExpressRoute circuit.

Screenshot of gateway connection bandwidth usage metric.

ExpressRoute Traffic Collector metrics

CPU Utilization - Split by instance

Aggregation type: Avg (of percentage of total utilized CPU)

Granularity: 5 min

You can view the CPU utilization of each ExpressRoute Traffic Collector instance. The CPU utilization may spike briefly during routine host maintenance, but prolonged high CPU utilization could indicate your ExpressRoute Traffic Collector is reaching a performance bottleneck.

Guidance: Set an alert for when avg CPU utilization exceeds a certain threshold.

Screenshot of CPU usage for ExpressRoute Traffic Collector.

Memory Utilization - Split by instance

Aggregation type: Avg (of percentage of total utilized Memory)

Granularity: 5 min

You can view the memory utilization of each ExpressRoute Traffic Collector instance. Memory utilization may spike briefly during routine host maintenance, but prolonged high memory utilization could indicate your Azure Traffic Collector is reaching a performance bottleneck.

Guidance: Set an alert for when avg memory utilization exceeds a certain threshold.

Screenshot of memory usage for ExpressRoute Traffic Collector.

Count of flow records processed - Split by instances or ExpressRoute circuit

Aggregation type: Count

Granularity: 5 min

You can view the count of number of flow records processed by ExpressRoute Traffic Collector, aggregated across ExpressRoute Circuits. Customer can split the metrics across each ExpressRoute Traffic Collector instance or ExpressRoute circuit when multiple circuits are associated to the ExpressRoute Traffic Collector. Monitoring this metric will help you understand if you need to deploy more ExpressRoute Traffic Collector instances or migrate ExpressRoute circuit association from one ExpressRoute Traffic Collector deployment to another.

Guidance: Splitting by circuits is recommended when multiple ExpressRoute circuits are associated with an ExpressRoute Traffic Collector deployment. This will help determine the flow count of each ExpressRoute circuit and ExpressRoute Traffic Collector utilization by each ExpressRoute circuit.

Screenshot of average flow records for an ExpressRoute circuit.

Alerts for ExpressRoute gateway connections

  1. To set up alerts, go to Azure Monitor, then select Alerts.

    alerts

  2. Select +Select Target and select the ExpressRoute gateway connection resource.

    target

  3. Define the alert details.

    action group

  4. Define and add the action group.

    add action group

Alerts based on each peering

each peering

Set up alerts for activity logs on circuits

In the Alert Criteria, you can select Activity Log for the Signal Type and select the Signal.

activity logs

More metrics in Log Analytics

You can also view ExpressRoute metrics by going to your ExpressRoute circuit resource and selecting the Logs tab. For any metrics you query, the output will contain the columns below.

Column Type Description
TimeGrain string PT1M (metric values are pushed every minute)
Count real Usually equal to 2 (each MSEE pushes a single metric value every minute)
Minimum real The minimum of the two metric values pushed by the two MSEEs
Maximum real The maximum of the two metric values pushed by the two MSEEs
Average real Equal to (Minimum + Maximum)/2
Total real Sum of the two metric values from both MSEEs (the main value to focus on for the metric queried)

Next steps

Set up your ExpressRoute connection.