Meraki Assurance Overview Page

Individual Experience Metric Thresholds

In addition to health thresholds, each experience metric has their own threshold based on the specific metric it is measuring.

Current Metrics and their respective thresholds are as follows:  

• For successful connects, clients that attempt a connection can either pass or fail. Each client pass and failure is averaged for the selected time period, and then averaged again across all the network's clients.
• For time to connect, clients are expected to connect to the network in 5 seconds or less. If the client fails to connect in 5 seconds or less, it will be marked as a failure and negatively impact this metric.
• For channel availability, a client that is connected to an AP under 70% overall utilization and less than 40% combined non-Wi-Fi and co-channel interference is considered a healthy client. If a client is connected to an AP over 70% overall utilization and more than 40% combined non-Wi-Fi and co-channel interference, this metric will be negatively impacted.
• For coverage, a client that has a total signal-to-noise ratio (SNR) at or above 20 dB has adequate coverage. If the clients SNR is lower than 20dB, this metric will be negatively impacted.

Wireless successful connects

Wireless successful connects measure the rate at which wireless clients successfully establish fresh connections across association, authentication, IP assignment, and DNS resolution phases. 

• Association - A device fails to associate if there is a mismatch between the device's supported 802.11 capabilities and the access point’s configurations, or if the access point (AP) has reached its maximum configured client limit. 
  • Sub-contributors:
    • Block listed client 
    • Feature conflict 
    • AP overloaded 
    • Roaming failures 
    • Cipher mismatch 
    • Other 
• Authentication - Failure occurs when the device provides incorrect security credentials, such as a wrong Pre-Shared Key (PSK) or invalid 802.1X/EAP certificates, preventing the establishment of a secure session. 
  • Sub-contributors:
    • Handshake timeout 
    • Invalid credentials 
    • RADIUS server issues 
    • AAA config errors 
    • Other 
• DHCP 
  • Sub-contributors:
    • Slow response 
    • Client not responding 
    • NAK or denial 
    • No DHCP response 
    • Proxy/relay config 
• DNS 
  • Sub-contributors:
    • DNS server unreachable 
    • Slow response 

Time to connect

Time to connect measures wireless client onboarding latency across association, authentication, IP assignment, and DNS resolution phases. Wireless clients should successfully establish fresh connections to the network in 5 seconds or less. 

• Association - A device fails to associate if there is a mismatch between the device's supported 802.11 capabilities and the access point’s configurations, or if the access point (AP) has reached its maximum configured client limit. 
• Authentication - Failure occurs when the device provides incorrect security credentials, such as a wrong Pre-Shared Key (PSK) or invalid 802.1X/EAP certificates, preventing the establishment of a secure session. 
• DHCP 
• DNS 

Channel availability

Channel availability measures RF channel capacity to determine whether wireless clients can reliably pass traffic over shared airtime. A wireless client has sufficient channel availability if its AP radio operates on a channel with less than 70% overall utilization and less than 40% combined non-Wi-Fi and co-channel Wi-Fi interference. 

• Co-Channel interference - A client fails when multiple access points or devices operating on the same frequency channel compete for airtime, increasing contention and raising the noise floor, which degrades the effective signal quality. 
  • Sub-contributors:
    • Out-of-network 
    • In-network 
• Non-Wi-Fi Interference - Clients will fail when when radio frequency from non 802.11 sources overlap with Wi-Fi channels, creating noise which prevents clients from successfully decoding data packets. Common non-802.11 sources include microwaves, Bluetooth devices, and motion sensors. 
• High Traffic - A client will fail when the volume of data being transmitted by all devices on a channel exceeds the available airtime, leading to excessing packet collisions, latency, and dropped connections from network congestion. 
  • Sub-contributors:
    • High client count 
    • High Bandwidth clients 

Coverage

Coverage measures the wireless signal quality of connected clients across the network. A wireless client has adequate coverage if its measured signal-to-noise ratio (SNR) is 20dB or greater. 

• Transient weak client signal - A client fails when its own transmit power is insuffcient to reliably reach the AP, leading to an asymmetric link where the client can hear the AP, but the AP cannot hear the client. This is often caused by a client’s physical distance to an AP. 
• Insufficient AP density - Failure occurs when there are too few APs installed in a physical space, creating "dead zones" or gaps in the radio frequency (RF) environment where signal strength drops below usable levels. 
• Admin power restriction - A client fails when the AP’s transmit power is manually capped or restricted by an administrator to a level that is too low to penetrate obstructions or cover the intended floor area, resulting in poor SNR at the cell edge. 
• Sticky client - A device that persistently remains connected to an underperforming access point despite the availability of better signal options, resulting in a consistent degradation of the user experience. 

Failure rate measures how frequent the individual contributing line item within the tab the user is viewing (e.g. SSID, band, etc.) is failing relative to itself. This indicates whether failures are isolated or are more broadly distributed. 

Selecting any of the above metrics will allow you to drill into the metric further though Insights view or Investigation view. 

Insights view provides a summary of the selected metric’s network impact as well as visual representation of success and the contributor failures. Additionally, it will intelligently highlight the fault domain(s) for the experience metric, helping guide users in troubleshooting efforts. 

The Investigation view lays out associated data related to the selected experience metric to identify top servers, access points, clients, and other information impacted by the failures.  

Requirements

• Successful onboarding and installation of the ThousandEyes agent on the Meraki MX. For more details, please refer to our Solution Requirements section in the Meraki MX ThousandEyes configuration guide.
• SDW+ Licensing

Score breakdown

Blue Score: This score is the ThousandEyes-based Application score. It’s shown here without modification besides the needed aggregation and averaging to ensure that it contains all the measurements from the selected time period. Metrics that are used to calculate this score are: Packet Loss, Latency, Jitter and Application Availability (any HTTP response code from the Application greater than 400 results in the App being unavailable). 

Red Score: Also known as the Customer Experience Score or Impact Score, is a score calculated based on the ThousandEyes score plus a weight that accounts for the number of potentially impacted clients. The idea is to summarize the Applications’ health and performance across the network and the impact of the outage on their users. This aggregation is also done based on the selected time period.

Points change: Show how the score has changed when compared to the previous time period. For example, if the data requested is for the last 2 hours it will show how the score has changed from the previous period of 2 hours. 

First Drawer: Network Layers

The goal of the first drawer is to highlight the potential network layers where the issue or outage might exist. Currently, they are divided into two:

• Internet Layer: It focuses on network issues that are present from the location where the ThousandEyes Enterprise agent is located up to the Application server, meaning anything across the Internet. The metrics that we capture are Packet Loss and Latency across the whole path.
• Application Layer: This layer showcases any issues with the application HTTP server itself, meaning any unexpected HTTP response or undesired response time. The metrics captured are Availability (Any HTTP response code > 400 is considered unavailable) and Response Time.

All the other metrics in the drawer are self-explanatory.

Second Drawer: Layer’s detail metrics

The second drawer provides detailed information on the value of the metrics that are captured for each layer and current or previous alerts that caused an outage or a performance degradation event. For example, for the Internet card, we see the Loss and Latency metrics and their values during the selected time period:

Most of the information on these pages is self-explanatory, but something to highlight is that on the line graphs, if the background is red, an alert was active during that time. We also provide a list of the start and end times for these alerts or issues at the bottom of the drawer. If an ongoing issue is detected, its end time will be shown as Active.

The second drawer will always include any information on triggered alerts that result in a performance issue for the network clients. If multiple alerts occur during that time and some are resolved, they will be ordered from latest to oldest, with the active alert always at the top. The third and final drawer can be accessed by clicking on one of the alerts.

Third Drawer: Client Impact and ThousandEyes Redirection

The third and final drawer focuses on showing the impacted clients by the previously selected alert. It can be accessed by clicking on one of the Alerts shown on the second drawer table.

This drawer provides information on which clients were potentially affected by the outage selected previously. It also allows redirection to the ThousandEyes Dashboard, where more information on the issue can be seen if desired. Please keep in mind that the ThousandEyes portal will require your login information if you have not already logged in.