Network Performance Metrics
This document was translated by ChatGPT
#1. Service List
Without inserting any code into the application, DeepFlow automatically generates all network performance metrics for services.
Database table name: flow_metrics.network.
#1.1 Tags
List of automatically injected tags: IP, protocol, port, collection location, cloud resources, K8s resources, K8s custom labels. Detailed field descriptions are as follows.
| Name | DisplayName | Description |
|---|---|---|
| time | Time | |
| region | Region | |
| az | Availability Zone | |
| host | VM Hypervisor | Host running virtual machine. |
| chost | Cloud Host | Including virtual machines |
| vpc | VPC | |
| subnet | Subnet | |
| router | Router | |
| dhcpgw | DHCP Gateway | |
| lb | Load Balancer | |
| lb_listener | Load Balancer Listener | |
| natgw | NAT Gateway | |
| redis | Redis | |
| rds | RDS | |
| pod_cluster | K8s Cluster | |
| pod_ns | K8s Namespace | |
| pod_node | K8s Node | |
| pod_ingress | K8s Ingress | |
| pod_service | K8s Service | |
| pod_group_type | K8s Workload Type | |
| pod_group | K8s Workload | Such as Deployment |
| pod | K8s POD | |
| service | Service | Deprecated,please use pod_service |
| auto_instance_type | Auto Instance Type | The type of 'auto_instance'. |
| auto_instance | Auto Instance Tag | The instance of IP |
| auto_service_type | Auto Service Type | The type of 'auto_service'. |
| auto_service | Auto Service Tag | On the basis of 'auto_instance' |
| gprocess | Process | |
| host_ip | VM Hypervisor | The management IP address of VM Hypervisor. |
| host_hostname | VM Hypervisor | The hostname of VM Hypervisor. |
| chost_ip | Cloud Host | The primary IP address of Cloud Host. |
| chost_hostname | Cloud Host | The hostname of Cloud Host. |
| pod_node_ip | K8s Node | The primary IP address of K8s Node. |
| pod_node_hostname | K8s Node | The hostname of K8s Node. |
| k8s.label | K8s Label | |
| k8s.annotation | K8s Annotation | |
| k8s.env | K8s Env | |
| cloud.tag | Cloud Tag | |
| os.app | OS APP | |
| ip | IP Address | |
| is_ipv4 | IPv4 Flag | |
| protocol | Network Protocol | |
| server_port | Server Port | |
| tap | Traffic Access Point | Deprecated,please use capture_network_type. |
| capture_network_type | Network Location | The network location for capturing traffic uses a fixed value (Cloud Network) to represent intra-cloud traffic |
| vtap | DeepFlow Agent | Deprecated,please use agent. |
| agent | DeepFlow Agent | |
| signal_source | Signal Source | |
| role | Role |
generate from csv file: network.en
#1.2 Metrics
List of metrics: throughput, load, latency, TCP anomalies, retransmissions, zero window. Detailed field descriptions are as follows.
| Field | DisplayName | Unit | Description |
|---|---|---|---|
| byte | Byte | Byte | |
| byte_tx | Byte TX | Byte | |
| byte_rx | Byte RX | Byte | |
| packet | Packet | Packet | |
| packet_tx | Packet TX | Packet | |
| packet_rx | Packet RX | Packet | |
| l3_byte | L3 Payload | Byte | |
| l3_byte_tx | L3 Payload TX | Byte | |
| l3_byte_rx | L3 Payload RX | Byte | |
| bpp | Bytes per Packet | Byte | |
| bpp_tx | Bytes per Packet TX | Byte | |
| bpp_rx | Bytes per Packet RX | Byte | |
| new_flow | New Flow | Flow | |
| closed_flow | Closed Flow | Flow | |
| flow_load | Active Flow | Flow | |
| syn_count | SYN Packet | Packet | |
| synack_count | SYN-ACK Packet | Packet | |
| l4_byte | L4 Payload | Byte | |
| l4_byte_tx | L4 Payload TX | Byte | |
| l4_byte_rx | L4 Payload RX | Byte | |
| retrans_syn | SYN Retransmission | Packet | |
| retrans_synack | SYN-ACK Retransmission | Packet | |
| retrans | TCP Retransmission | Packet | |
| retrans_tx | TCP Client Retransmission | Packet | |
| retrans_rx | TCP Server Retransmission | Packet | |
| zero_win | TCP ZeroWindow | Packet | |
| zero_win_tx | TCP Client ZeroWindow | Packet | |
| zero_win_rx | TCP Server ZeroWindow | Packet | |
| retrans_syn_ratio | SYN Retrans. % | % | |
| retrans_synack_ratio | SYN-ACK Retrans. % | % | |
| retrans_ratio | TCP Retrans. % | % | |
| retrans_tx_ratio | TCP Client Retrans. % | % | |
| retrans_rx_ratio | TCP Server Retrans. % | % | |
| zero_win_ratio | TCP ZeroWindow % | % | |
| zero_win_tx_ratio | TCP Client ZeroWindow % | % | |
| zero_win_rx_ratio | TCP Server ZeroWindow % | % | |
| tcp_establish_fail | Error | Flow | |
| client_establish_fail | Client Error | Flow | |
| server_establish_fail | Server Error | Flow | |
| tcp_establish_fail_ratio | Error % | % | |
| client_establish_fail_ratio | Client Error % | % | |
| server_establish_fail_ratio | Client Error % | % | |
| tcp_transfer_fail | Transfer Error | Flow | All transfer errors. |
| tcp_transfer_fail_ratio | Transfer Error % | % | |
| tcp_rst_fail | RST | Flow | All RST errors. |
| tcp_rst_fail_ratio | RST % | % | |
| client_source_port_reuse | Est. - Client Port Reuse | Flow | |
| server_syn_miss | Est. - Server SYN Miss | Flow | |
| client_establish_other_rst | Est. - Client Other RST | Flow | |
| client_ack_miss | Est. - Client ACK Miss | Flow | |
| server_reset | Est. - Server Direct RST | Flow | |
| server_establish_other_rst | Est. - Server Other RST | Flow | |
| client_rst_flow | Transfer - Client RST | Flow | |
| server_rst_flow | Transfer - Server RST | Flow | |
| server_queue_lack | Transfer - Server Queue Overflow | Flow | |
| tcp_timeout | Transfer - TCP Timeout | Flow | |
| client_half_close_flow | Close - Client Half Close | Flow | |
| server_half_close_flow | Close - Server Half Close | Flow | |
| rtt | Avg TCP Est. Delay | us | |
| rtt_client | Avg TCP Est. Client Delay | us | |
| rtt_server | Avg TCP Est. Server Delay | us | |
| srt | Avg TCP/ICMP Response Delay | us | |
| art | Avg Data Delay | us | |
| cit | Avg Client Idle Delay | us | |
| rtt_max | Max TCP Est. Delay | us | |
| rtt_client_max | Max TCP Est. Client Delay | us | |
| rtt_server_max | Max TCP Est. Server Delay | us | |
| srt_max | Max TCP/ICMP Response Delay | us | |
| art_max | Max Data Delay | us | |
| cit_max | Max Client Idle Delay | us | |
| l7_request | Request | ||
| l7_response | Response | ||
| rrt | Avg App. Delay | us | |
| rrt_max | Max App. Delay | us | |
| l7_error | App. Error | ||
| l7_client_error | App. Client Error | ||
| l7_server_error | App. Server Error | ||
| l7_timeout | App. Server Timeout | ||
| l7_error_ratio | App. Error % | % | |
| l7_client_error_ratio | App. Client Error % | % | |
| l7_server_error_ratio | App. Server Error % | % | |
| row | Row Count |
generate from csv file: network.en
#1.3 Grafana Dashboard
Based on the above data, you can build rich dashboards through Grafana. We have pre-configured a Network - K8s Pod dashboard in Grafana, as shown below:

Network K8s Pod
You can also visit DeepFlow Online Demo (opens new window) to see the effect.
#2. Universal Service Map
Without inserting any code into the application, DeepFlow automatically generates the full-stack access paths and network performance metrics for all services.
Database table name: flow_metrics.network_map.
#2.1 Tags
List of automatically injected tags: IP, protocol, port, collection location, cloud resources, K8s resources, K8s custom labels. Detailed field descriptions are as follows.
| Name | DisplayName | Description |
|---|---|---|
| time | Time | |
| region | Region | |
| az | Availability Zone | |
| host | VM Hypervisor | Host running virtual machine. |
| chost | Cloud Host | Including virtual machines |
| vpc | VPC | |
| subnet | Subnet | |
| router | Router | |
| dhcpgw | DHCP Gateway | |
| lb | Load Balancer | |
| lb_listener | Load Balancer Listener | |
| natgw | NAT Gateway | |
| redis | Redis | |
| rds | RDS | |
| pod_cluster | K8s Cluster | |
| pod_ns | K8s Namespace | |
| pod_node | K8s Node | |
| pod_ingress | K8s Ingress | |
| pod_service | K8s Service | |
| pod_group_type | K8s Workload Type | |
| pod_group | K8s Workload | Such as Deployment |
| pod | K8s POD | |
| service | Service | Deprecated,please use pod_service |
| auto_instance_type | Auto Instance Type | The type of 'auto_instance'. |
| auto_instance | Auto Instance Tag | The instance of IP |
| auto_service_type | Auto Service Type | The type of 'auto_service'. |
| auto_service | Auto Service Tag | On the basis of 'auto_instance' |
| gprocess | Process | |
| tap_port_host | Tap Port Host | |
| tap_port_chost | Tap Port Cloud Host | |
| tap_port_pod_node | Tap Port K8s Node | |
| capture_nic_host | Host of Capture NIC | |
| capture_nic_chost | Cloud Host of Capture NIC | |
| capture_nic_pod_node | K8s Node of Capture NIC | |
| host_ip | VM Hypervisor | The management IP address of VM Hypervisor. |
| host_hostname | VM Hypervisor | The hostname of VM Hypervisor. |
| chost_ip | Cloud Host | The primary IP address of Cloud Host. |
| chost_hostname | Cloud Host | The hostname of Cloud Host. |
| pod_node_ip | K8s Node | The primary IP address of K8s Node. |
| pod_node_hostname | K8s Node | The hostname of K8s Node. |
| k8s.label | K8s Label | |
| k8s.annotation | K8s Annotation | |
| k8s.env | K8s Env | |
| cloud.tag | Cloud Tag | |
| os.app | OS APP | |
| ip | IP Address | |
| is_ipv4 | IPv4 Flag | |
| is_internet | Internet IP Flag | Whether the IP address is an external Internet address. |
| protocol | Network Protocol | |
| tunnel_type | Tunnel Type | |
| server_port | Server Port | |
| tap | Traffic Access Point | Traffic Access Point. Use a fixed value (virtual network) to represent cloud traffic |
| capture_network_type | Capture Network Type | Traffic Access Point. Use a fixed value (virtual network) to represent cloud traffic |
| vtap | DeepFlow Agent | |
| agent | DeepFlow Agent | |
| nat_source | NAT Source | |
| tap_port | TAP Port Identifier | When the value of tap_port_type is 'Local NIC' |
| tap_port_name | TAP Port Name | When the value of tap_port_type is 'Local NIC' |
| tap_port_type | TAP Port Type | Indicates the type of traffic collection location |
| capture_nic | Capture NIC ID | When the value of tap_port_type is 'Local NIC' |
| capture_nic_name | Capture NIC Name | When the value of tap_port_type is 'Local NIC' |
| capture_nic_type | Capture NIC Type | Indicates the type of traffic collection location |
| tap_side | TAP Side | The logical location of the collection location in the traffic path |
| observation_point | Observation Point | The logical location of the collection location in the traffic path |
| signal_source | Signal Source |
generate from csv file: network_map.en
#2.2 Metrics
List of metrics: throughput, load, latency, TCP anomalies, retransmissions, zero window. Detailed field descriptions are as follows.
| Field | DisplayName | Unit | Description |
|---|---|---|---|
| byte | Byte | Byte | |
| byte_tx | Byte TX | Byte | |
| byte_rx | Byte RX | Byte | |
| packet | Packet | Packet | |
| packet_tx | Packet TX | Packet | |
| packet_rx | Packet RX | Packet | |
| l3_byte | L3 Payload | Byte | |
| l3_byte_tx | L3 Payload TX | Byte | |
| l3_byte_rx | L3 Payload RX | Byte | |
| bpp | Bytes per Packet | Byte | |
| bpp_tx | Bytes per Packet TX | Byte | |
| bpp_rx | Bytes per Packet RX | Byte | |
| new_flow | New Flow | Flow | |
| closed_flow | Closed Flow | Flow | |
| flow_load | Active Flow | Flow | |
| syn_count | SYN Packet | Packet | |
| synack_count | SYN-ACK Packet | Packet | |
| l4_byte | L4 Payload | Byte | |
| l4_byte_tx | L4 Payload TX | Byte | |
| l4_byte_rx | L4 Payload RX | Byte | |
| direction_score | Direction Score | The higher the score | |
| retrans_syn | SYN Retransmission | Packet | |
| retrans_synack | SYN-ACK Retransmission | Packet | |
| retrans | TCP Retransmission | Packet | |
| retrans_tx | TCP Client Retransmission | Packet | |
| retrans_rx | TCP Server Retransmission | Packet | |
| zero_win | TCP ZeroWindow | Packet | |
| zero_win_tx | TCP Client ZeroWindow | Packet | |
| zero_win_rx | TCP Server ZeroWindow | Packet | |
| retrans_syn_ratio | SYN Retrans. % | % | |
| retrans_synack_ratio | SYN-ACK Retrans. % | % | |
| retrans_ratio | TCP Retrans. % | % | |
| retrans_tx_ratio | TCP Client Retrans. % | % | |
| retrans_rx_ratio | TCP Server Retrans. % | % | |
| zero_win_ratio | TCP ZeroWindow % | % | |
| zero_win_tx_ratio | TCP Client ZeroWindow % | % | |
| zero_win_rx_ratio | TCP Server ZeroWindow % | % | |
| tcp_establish_fail | Error | Flow | |
| client_establish_fail | Client Error | Flow | |
| server_establish_fail | Server Error | Flow | |
| tcp_establish_fail_ratio | Error % | % | |
| client_establish_fail_ratio | Client Error % | % | |
| server_establish_fail_ratio | Client Error % | % | |
| tcp_transfer_fail | Transfer Error | Flow | All transfer errors. |
| tcp_transfer_fail_ratio | Transfer Error % | % | |
| tcp_rst_fail | RST | Flow | All RST errors. |
| tcp_rst_fail_ratio | RST % | % | |
| client_source_port_reuse | Est. - Client Port Reuse | Flow | |
| server_syn_miss | Est. - Server SYN Miss | Flow | |
| client_establish_other_rst | Est. - Client Other RST | Flow | |
| client_ack_miss | Est. - Client ACK Miss | Flow | |
| server_reset | Est. - Server Direct RST | Flow | |
| server_establish_other_rst | Est. - Server Other RST | Flow | |
| client_rst_flow | Transfer - Client RST | Flow | |
| server_rst_flow | Transfer - Server RST | Flow | |
| server_queue_lack | Transfer - Server Queue Overflow | Flow | |
| tcp_timeout | Transfer - TCP Timeout | Flow | |
| client_half_close_flow | Close - Client Half Close | Flow | |
| server_half_close_flow | Close - Server Half Close | Flow | |
| rtt | Avg TCP Est. Delay | us | |
| rtt_client | Avg TCP Est. Client Delay | us | |
| rtt_server | Avg TCP Est. Server Delay | us | |
| srt | Avg TCP/ICMP ACK Delay | us | |
| art | Avg Data Delay | us | |
| cit | Avg Client Idle Delay | us | |
| rtt_max | Max TCP Est. Delay | us | |
| rtt_client_max | Max TCP Est. Client Delay | us | |
| rtt_server_max | Max TCP Est. Server Delay | us | |
| srt_max | Max TCP/ICMP ACK Delay | us | |
| art_max | Max Data Delay | us | |
| cit_max | Max Client Idle Delay | us | |
| l7_request | Request | ||
| l7_response | Response | ||
| rrt | Avg App. Delay | us | |
| rrt_max | Max App. Delay | us | |
| l7_error | App. Error | ||
| l7_client_error | App. Client Error | ||
| l7_server_error | App. Server Error | ||
| l7_timeout | App. Server Timeout | ||
| l7_error_ratio | App. Error % | % | |
| l7_client_error_ratio | App. Client Error % | % | |
| l7_server_error_ratio | App. Server Error % | % | |
| row | Row Count |
generate from csv file: network_map.en
#2.3 Grafana Dashboard
Based on the above data, you can build rich dashboards through Grafana. We have pre-configured a Network - K8s Pod Map dashboard in Grafana, as shown below:

Network K8s Pod Map
You can also visit DeepFlow Online Demo (opens new window) to see the effect.