Download page Network Dashboard.
The Network Dashboard provides a network-layer view of your application where you can quickly determine if any part of the network is impacting application performance. The bottom charts show the overall network health. Each tier and link shows network KPI (Key Performance Indicator) metrics. Network Agents detect intermediate load balancers and TCP endpoints. You can:
- Enable baselining to visualize network elements with performance issues (red/yellow links and tiers)
- View network metric charts in tier/link popups
- Find root causes in a specific network link. Right-click and select View Metrics to open a context-sensitive dashboard for the element.
- The tier dashboard shows correlations between outlier (Slow/Very Slow/Stalled/Error) transactions and network-performance metrics
- The link dashboard shows network-performance issues on the client tier, server tier, and network path
- Find root causes in a specific connection. When you narrow the issue down to a specific network link, you can:
- Configure the Network Agents to collect detailed metrics for the individual Connections used on that link.
- Drill down to find the root cause in an individual Connection.
Copy IP addresses and ports for individual connections from the Network Dashboard enabling you to forward this information to operations and other personnel when troubleshooting an issue. To copy IP addresses and ports:
- Click a link to open the link popup, and go to the Connections tab.
- Select the connections of interest (use Ctrl-click to select multiple connections).
- Right-click the selection and select Copy IP addresses to clipboard.
- In many cases, the Application Dashboard and Network Dashboard might show different topologies for the same application. This is the expected behavior. Unlike App Server Agents, Network Agents detect intermediate network devices between nodes, and consider these devices when collecting metrics. The two dashboards show the same application but from different perspectives: an application perspective and a multi-layer, network/application perspective.
- This release does not support these Network Flow Map features:
- Federated Flow Maps
- Visualization of flows between web servers and APM entry tiers
- The Network Flow Map does not filter out connections based on the selected time range
- In some cases, the KPIs for a tier or link might be different in the Network tab versus the Flow Map. The popup shows the latest data; the Flow Map updates data every two minutes. Any discrepancy between KPI values is due to this difference in reporting times.
- If you open the Network Flow Map for an individual node, the KPI metrics for node-to-load-balancer and node-to-tcp-endpoint links show KPIs for all nodes in the parent tier (instead of KPIs for the individual node only). To view KPIs for the individual node, open the link popup and review the Connection KPIs.
Network Dashboard Reference
The Network Dashboard shows:
- Network-based key performance Indicators (KPIs) for the monitored nodes in each tier:
- Average Latency (#ms) for individual packets
- Number of Performance Impacting Events (PIE)
- Number of Errors
- KPIs for all network links. Click a load balancer or a network element to review the set of five associated Connections. (A gray link indicates a connection that a Network Agent detected but did not directly monitor.) If a link or load balancer is associated with more connections, click See all connections. The Connections page lists the details of all the connections.
- The Network Dashboard shows a TCP endpoint icon when a Network Agent observes traffic between a monitored node and an endpoint (IP address, TCP port, and protocol), but cannot determine if the endpoint represents:
- An application node that sends or receives traffic for the application, or
- An intermediate device (proxy or load balancer) that transfers traffic to and from another node over a separate connection
- Network Agents automatically detect load balancers, which often mask internal IP addresses and have the effect of splitting an in-flight application message into two separate TCP connections. The auto-detection of TCP endpoints and load balancers between application nodes enable you to easily identify the exact locations where network issues are occurring.
- Links are colored in the Network Dashboard as:
- Green/yellow/red, to visualize the network performance compared to the overall performance baseline for that link
- Blue, if baselining is disabled
- Gray, for connections between application nodes and TCP endpoints. In this example, the gray link indicates "The connected node and endpoint are either the same device, or two devices exchanging traffic over a separate, unmonitored connection."
- The Network Dashboard also shows KPIs
for the entire application. You can review:
- Overall throughput that the application places on the network
- Rate of network packet loss (packets sent but never received)
- Number of Performance Impacting Events (PIE), which indicate potential problems on one or more nodes or connections
- Number of errors for the entire application