在Catalyst 9800上使用Telegraf、ConsumeDB和Grafana配置高級gRPC工作流程

簡介

本文檔介紹如何部署Telegraf、ConsumeDB和Grafana (TIG)堆疊以及如何將其與Catalyst 9800互聯。

必要條件

本文檔透過複雜的整合演示了Catalyst 9800的程式設計介面功能。本文檔旨在展示如何根據任何需求完全定製這些功能，以及如何節省日常時間。此處展示的部署依賴於gRPC，並提供遙測配置，使來自任何Telegraf、ConsumeDB、Grafana (TIG)可觀察堆疊中的Catalyst 9800無線資料可用。

需求

思科建議您瞭解以下主題：

• Catalyst Wireless 9800配置型號。 
• 網路可程式設計性和資料模型。
• TIG堆疊基礎知識。

採用元件

本文中的資訊係根據以下軟體和硬體版本：

• Catalyst 9800-CL (v. 17.12.03)。
• Ubuntu (v. 22.04.03)。
• InfloodDB (v. 1.06.07)。
• Telegraf (v. 1.21.04)。
• 格拉法納(v. 10.02.01)。

本文中的資訊是根據特定實驗室環境內的裝置所建立。文中使用到的所有裝置皆從已清除（預設）的組態來啟動。如果您的網路運作中，請確保您瞭解任何指令可能造成的影響。

設定

網路圖表

組態

在本示例中，在9800-CL上使用gRPC撥出來配置遙測，以將儲存資訊的Telegraf應用程式上的資訊推送到ConsumeDB資料庫中。這裡使用了兩個裝置，

• 託管整個TIG堆疊的Ubuntu伺服器。
• Catalyst 9800-CL。

本配置指南並不關注這些裝置的整個部署，而是關注每個應用程式上要正確傳送、接收和顯示9800資訊所需的配置。

步驟 1.準備資料庫

進入配置部分之前，請確保您的Confusion例項運行正常。使用systemctl status命令，如果您使用的是Linux發行版，這可以很容易地完成。

admin@tig:~$ systemctl status influxd
● influxdb.service - InfluxDB is an open-source, distributed, time series database
     Loaded: loaded (/lib/systemd/system/influxdb.service; enabled; vendor preset: enabled)
     Active: active (running) since Wed 2023-06-14 13:06:18 UTC; 2 weeks 5 days ago
       Docs: https://docs.influxdata.com/influxdb/
   Main PID: 733 (influxd)
      Tasks: 15 (limit: 19180)
     Memory: 4.2G
        CPU: 1h 28min 47.366s
     CGroup: /system.slice/influxdb.service
             └─733 /usr/bin/influxd -config /etc/influxdb/influxdb.conf

例如，Telegraf需要一個資料庫來儲存測量結果，以及一個使用者來連線至此測量結果。可以使用以下命令，從InfloodDB CLI輕鬆建立這些命令：

admin@tig:~$ influx
Connected to http://localhost:8086 version 1.8.10
InfluxDB shell version: 1.8.10
> create database TELEGRAF
> create user telegraf with password 'YOUR_PASSWORD'

現在已建立資料庫，Telegraf可以設定成將測量結果正確儲存到資料庫中。

步驟 2.準備電報

此示例中只有兩個有趣的Telegraf配置起作用。這些可由配置檔案/etc/telegraf/telegraf.conf建立（對於運行在Unix上的應用程式，通常如此）。

第一個命令會宣告Telegraf所使用的輸出。如上所述，此處使用ConfusionDB，並在telegraf.conf檔案的輸出部分中配置如下：

###############################################################################
#                            OUTPUT PLUGINS                                   #
###############################################################################
# Output Plugin InfluxDB
[[outputs.influxdb]]
## The full HTTP or UDP URL for your InfluxDB instance.
#   ##
#   ## Multiple URLs can be specified for a single cluster, only ONE of the
#   ## urls will be written to each interval.
  urls = [ "http://127.0.0.1:8086" ]
#   ## The target database for metrics; will be created as needed.
#   ## For UDP url endpoint database needs to be configured on server side.
  database = "TELEGRAF"
#   ## HTTP Basic Auth
  username = "telegraf"
  password = "YOUR_PASSWORD"

這指示Telegraf進程將其接收到的資料儲存在埠8086上運行於同一主機上的ConsumeDB中，並使用名為「TELEGRAF」的資料庫（以及訪問它的憑證telegraf/YOUR_PASSWORD）。

如果第一個宣告的是輸出格式，第二個當然是輸入格式。要通知Telegraf它接收的資料來自使用遙測的思科裝置，您可以使用cisco_telemetry_mdt」輸入模組。要配置此功能，您只需在/etc/telegraf/telegraf.conf檔案中增加以下行：

###############################################################################
#                             INPUT PLUGINS                                   #
###############################################################################
# # Cisco model-driven telemetry (MDT) input plugin for IOS XR, IOS XE and NX-OS platforms
[[inputs.cisco_telemetry_mdt]]
#  ## Telemetry transport can be "tcp" or "grpc".  TLS is only supported when
#  ## using the grpc transport.
    transport = "grpc"
#
#  ## Address and port to host telemetry listener
    service_address = ":57000"
#  ## Define aliases to map telemetry encoding paths to simple measurement names
[inputs.cisco_telemetry_mdt.aliases]
    ifstats = "ietf-interfaces:interfaces-state/interface/statistics"

這使得在主機上運行的Telegraf應用程式(在預設埠57000)能夠對來自WLC的接收資料進行解碼。

儲存配置後，請確保重新啟動Telegraf以將其應用於服務。還要確保服務已正確重新啟動：

admin@tig:~$ sudo systemctl restart telegraf
admin@tig:~$ systemctl status telegraf.service
● telegraf.service - Telegraf
     Loaded: loaded (/lib/systemd/system/telegraf.service; enabled; vendor preset: enabled)
     Active: active (running) since Mon 2023-07-03 17:12:49 UTC; 2min 18s ago
       Docs: https://github.com/influxdata/telegraf
   Main PID: 110182 (telegraf)
      Tasks: 10 (limit: 19180)
     Memory: 47.6M
        CPU: 614ms
     CGroup: /system.slice/telegraf.service
             └─110182 /usr/bin/telegraf -config /etc/telegraf/telegraf.conf -config-directory /etc/telegraf/telegraf.d

步驟 3.確定包含所需度量的遙測訂閱

如上所述，在思科裝置上和其他許多裝置上，度量是按照YANG模型組織的。可以在此處找到每個版本的IOS XE（用於9800）的特定思科YANG型號，特別是本示例中使用的用於都柏林IOS XE 17.12.03的型號。

在本例中，我們側重於從使用的9800-CL例項收集CPU利用率指標。透過檢查Cisco IOS XE都柏林17.12.03的YANG型號，可以確定哪個模組包含控制器的CPU使用率，尤其是最近5秒內的使用率。這些是Cisco-IOS-XE-process-cpu-oper模組的一部分，位於cpu-utilization分組（枝葉5秒）下。

步驟 4.在控制器上啟用NETCONF

gRPC撥出架構依賴NETCONF來正常運行。因此，必須在9800上啟用此功能，可透過運行以下命令實現此功能：

WLC(config)#netconf ssh
WLC(config)#netconf-yang

步驟 5.在控制器上配置遙測訂閱

根據YANG模型確定的度量的XPaths(a.k.a，XML路徑語言)之後，可以從9800 CLI輕鬆配置遙測訂閱，以便開始將這些遙測訂閱流式傳輸到步驟2中配置的Telegraf例項。這可以透過執行以下命令來完成：

WLC(config)#telemetry ietf subscription 101
WLC(config-mdt-subs)#encoding encode-kvgpb
WLC(config-mdt-subs)#filter xpath /process-cpu-ios-xe-oper:cpu-usage/cpu-utilization/five-seconds
WLC(config-mdt-subs)#source-address 10.48.39.130
WLC(config-mdt-subs)#stream yang-push
WLC(config-mdt-subs)#update-policy periodic 100
WLC(config-mdt-subs)#receiver ip address 10.48.39.98 57000 protocol grpc-tcp

在此代碼塊中，首先定義識別符號為101的遙測訂閱。訂用識別符號可以是<0-2147483647>之間的任何數字，只要它不與其他訂用重疊。此訂閱已按以下順序進行配置：

• 使用的編碼方法，在使用gRPC傳輸協定時，必須是kvGPB。
• 訂用所傳送度量的過濾器，是定義我們感興趣度量的XPath(若要知道，請使用/process-cpu-ios-xe-oper:cpu-usage/cpu-utilization/five-seconds)。
• 控制器用於傳送度量的源IP地址。
• 用於傳遞度量的流型別，在本例中為YANG Push IETF標準。
• 控制器用來在100^秒內傳送資料給訂戶的頻率。在本例中，它被配置為每秒定期傳送更新。
• 接收器IP地址和埠號，以及用於在控制器和使用者之間通訊的協定。在本示例中，gRPC-TCP用於將度量傳送到埠57000上的主機10.48.39.98。

步驟 6.配置Grafana資料來源

現在，控制器開始向Telegraf傳送資料，這些資料儲存在TELEGRAF HonglumDB資料庫中，現在應該配置Grafana使其瀏覽這些指標。

在您的Grafana GUI中，導航到Home > Connections > Connect data，然後使用搜尋欄查詢InflostDB資料來源。

選擇此資料來源型別，然後使用「建立ConsumeDB資料來源」按鈕連線Grafana和在步驟1中建立的TELEGRAPH資料庫。

填入熒幕上顯示的表單，特別是提供：

• 資料來源的名稱。
• 所使用的ConfusionDB例項的URL。
• 使用的資料庫名稱（在本範例中為「TELEGRAF」）。
• 使用者定義的認證以存取它（在此範例中為telegraf/YOUR_PASSWORD）。

步驟 7.建立儀表板

Grafana視覺化內容被組織到控制台中。要建立包含Catalyst 9800度量視覺化的儀表板，請導航到首頁>儀表板，然後使用「新建儀表板」按鈕

這樣會開啟新建的圖示板。按一下齒輪圖示以存取圖示板引數並變更其名稱。本例使用「Catalyst 9800遙測」。執行此操作後，請使用「儲存控制台」按鈕儲存控制台。

步驟 8.將視覺化新增至儀表板

既然資料已經正確傳送、接收和儲存，而且Grafana可以訪問這個儲存位置，那麼現在是時候為他們建立視覺化了。

從任何Grafana儀表板中，使用「新增」按鈕，並從顯示的功能表中選取「視覺化」，建立測量結果的視覺化。

 

這將打開所建立的視覺化的編輯面板：

在此面板中，選取

• 您在步驟6中建立的資料來源的名稱，在本示例中為TELEGRAF。
• 包含要視覺化資料的度量（架構），在本例中為「Cisco-IOS-XE-process-cpu-oper：cpu-usage/cpu-utilization」。
• 代表您要視覺化之測量結果的資料庫欄位，在本範例中為「five_seconds」。
• 此範例中視覺化功能的標題「CPU使用率9800-CL」。

按下上一個圖中的「儲存/應用」按鈕後，顯示Catalyst 9800控制器隨時間變化的CPU使用情況的視覺化內容將增加到控制台中。使用磁片圖示按鈕可以儲存對控制台所做的更改。

驗證

WLC運行配置

Building configuration...
Current configuration : 112215 bytes
!
! Last configuration change at 14:28:36 UTC Thu May 23 2024 by admin
! NVRAM config last updated at 14:28:23 UTC Thu May 23 2024 by admin
!
version 17.12
[...]
aaa new-model
!
!
aaa authentication login default local
aaa authentication login local-auth local
aaa authentication dot1x default group radius
aaa authorization exec default local 
aaa authorization network default group radius 
[...]
vlan internal allocation policy ascending
!
vlan 39 
!
vlan 1413
 name VLAN_1413
!
!
interface GigabitEthernet1
 switchport access vlan 1413
 negotiation auto
 no mop enabled
 no mop sysid
!
interface GigabitEthernet2
 switchport trunk allowed vlan 39,1413
 switchport mode trunk
 negotiation auto
 no mop enabled
 no mop sysid
!
interface Vlan1
 no ip address
 no ip proxy-arp
 no mop enabled
 no mop sysid
!
interface Vlan39
 ip address 10.48.39.130 255.255.255.0
 no ip proxy-arp
 no mop enabled
 no mop sysid
[...]
telemetry ietf subscription 101
 encoding encode-kvgpb
 filter xpath /process-cpu-ios-xe-oper:cpu-usage/cpu-utilization
 source-address 10.48.39.130
 stream yang-push
 update-policy periodic 1000
 receiver ip address 10.48.39.98 57000 protocol grpc-tcp
[...]
netconf-yang

Telegraf配置

# Configuration for telegraf agent
[agent]
  metric_buffer_limit = 10000
  collection_jitter = "0s"
  debug = true
  quiet = false
  flush_jitter = "0s"
  hostname = ""
  omit_hostname = false

###############################################################################
#                            OUTPUT PLUGINS                                   #
###############################################################################
# Configuration for sending metrics to InfluxDB
[[outputs.influxdb]]
  urls = ["http://127.0.0.1:8086"]
  database = "TELEGRAF"
  username = "telegraf"
  password = "Wireless123#"

###############################################################################
#                            INPUT PLUGINS                                    #
###############################################################################

###############################################################################
#                            SERVICE INPUT PLUGINS                            #
###############################################################################
# # Cisco model-driven telemetry (MDT) input plugin for IOS XR, IOS XE and NX-OS platforms
[[inputs.cisco_telemetry_mdt]]
  transport = "grpc"
  service_address = "10.48.39.98:57000"
 [inputs.cisco_telemetry_mdt.aliases]
    ifstats = "ietf-interfaces:interfaces-state/interface/statistics"

GrowthDB配置

### Welcome to the InfluxDB configuration file.
reporting-enabled = false
[meta]
  dir = "/var/lib/influxdb/meta"

[data]
  dir = "/var/lib/influxdb/data"
  wal-dir = "/var/lib/influxdb/wal"

[retention]
  enabled = true
  check-interval = "30m"

Grafana配置

#################################### Server ####################################
[server]
http_addr = 10.48.39.98
domain = 10.48.39.98

疑難排解

WLC One Stop-Shop反射

從WLC端，首先要檢驗的是與程式設計介面相關的進程是否已啟動並正在運行。

#show platform software yang-management process
confd : Running
nesd : Running
syncfd : Running
ncsshd : Running               <-- NETCONF / gRPC Dial-Out
dmiauthd : Running             <-- For all of them, Device Managment Interface needs to be up.  
nginx : Running                <-- RESTCONF 
ndbmand : Running
pubd : Running
gnmib : Running                <-- gNMI 

對於NETCONF（由gRPC撥出），這些命令還可以幫助檢查進程的狀態。

WLC#show netconf-yang status 
netconf-yang: enabled
netconf-yang candidate-datastore: disabled
netconf-yang side-effect-sync: enabled
netconf-yang ssh port: 830
netconf-yang turbocli: disabled
netconf-yang ssh hostkey algorithms: rsa-sha2-256,rsa-sha2-512,ssh-rsa
netconf-yang ssh encryption algorithms: aes128-ctr,aes192-ctr,aes256-ctr,aes128-cbc,aes256-cbc
netconf-yang ssh MAC algorithms: hmac-sha2-256,hmac-sha2-512,hmac-sha1
netconf-yang ssh KEX algorithms: diffie-hellman-group14-sha1,diffie-hellman-group14-sha256,ecdh-sha2-nistp256,ecdh-sha2-nistp384,ecdh-sha2-nistp521,diffie-hellman-group16-sha512

檢查進程狀態後，另一個重要檢查是Catalyst 9800和電信接收器之間的遙測連線狀態。可使用「show telemetry connection all」命令檢視它。

WLC#show telemetry connection all 
Telemetry connections

Index Peer Address               Port  VRF Source Address             State      State Description
----- -------------------------- ----- --- -------------------------- ---------- --------------------
28851 10.48.39.98                57000 0   10.48.39.130               Active     Connection up       

如果WLC和接收器之間的遙測連線啟動，您還可以使用show telemetry ietf subscription all brief命令確保配置的訂閱有效。

WLC#show telemetry ietf subscription all brief
ID         Type       State      State Description
101        Configured Valid      Subscription validated

此命令的詳細版本show telemetry ietf subscription all detail提供有關訂閱的更多資訊，有助於指出其配置中的問題。

WLC#show telemetry ietf subscription all detail 
Telemetry subscription detail:

  Subscription ID: 101
  Type: Configured
  State: Valid
  Stream: yang-push
  Filter:
    Filter type: xpath
    XPath: /process-cpu-ios-xe-oper:cpu-usage/cpu-utilization
  Update policy:
    Update Trigger: periodic
    Period: 1000
  Encoding: encode-kvgpb
  Source VRF: 
  Source Address: 10.48.39.130
  Notes: Subscription validated

  Named Receivers:
    Name                                              Last State Change  State                 Explanation                                                   
    -------------------------------------------------------------------------------------------------------------------------------------------------------
    grpc-tcp://10.48.39.98:57000                      05/23/24 08:00:25  Connected                                                                           

確認網路連線能力

Catalyst 9800控制器將gRPC資料傳送到為每個遙測訂閱配置的接收器埠。

WLC#show run | include receiver ip address
receiver ip address 10.48.39.98 57000 protocol grpc-tcp

要驗證WLC和此已配置埠上的接收器之間的網路連線，可使用多種工具。

從WLC中，可以在已設定的接收器IP/連線埠(這裡為10.48.39.98：57000)上使用telnet來驗證此接收器是否已開啟以及是否可從控制器本身連線。如果沒有封鎖流量，則連線埠必須在輸出中顯示為開啟：

WLC#telnet 10.48.39.98 57000
Trying 10.48.39.98, 57000 ... Open <-------

或者，您可以使用來自任何主機的Nmap來確保接收器已在配置的埠上正確曝光。

$ sudo nmap -sU -p 57000 10.48.39.98
Starting Nmap 7.95 ( https://nmap.org ) at 2024-05-17 13:12 CEST
Nmap scan report for air-1852e-i-1.cisco.com (10.48.39.98)
Host is up (0.020s latency).

PORT      STATE         SERVICE
57000/udp open|filtered unknown

Nmap done: 1 IP address (1 host up) scanned in 0.35 seconds

記錄與除錯

2024/05/23 14:40:36.566486156 {pubd_R0-0}{2}: [mdt-ctrl] [30214]: (note): **** Event Entry: Configured legacy receiver creation/modification of subscription 101 receiver 'grpc-tcp://10.48.39.98:57000'
2024/05/23 14:40:36.566598609 {pubd_R0-0}{2}: [mdt-ctrl] [30214]: (note): Use count for named receiver 'grpc-tcp://10.48.39.98:57000' set to 46.
2024/05/23 14:40:36.566600301 {pubd_R0-0}{2}: [mdt-ctrl] [30214]: (note): {subscription receiver event='configuration created'} received for subscription 101 receiver 'grpc-tcp://10.48.39.98:57000'
[...]
2024/05/23 14:40:36.572402901 {pubd_R0-0}{2}: [pubd] [30214]: (info): Collated data collector filters for subscription 101.
2024/05/23 14:40:36.572405081 {pubd_R0-0}{2}: [pubd] [30214]: (debug): Creating periodic sensor for subscription 101.
2024/05/23 14:40:36.572670046 {pubd_R0-0}{2}: [pubd] [30214]: (info): Creating data collector type 'ei_do periodic' for subscription 101 using filter '/process-cpu-ios-xe-oper:cpu-usage/cpu-utilization'.
2024/05/23 14:40:36.572670761 {pubd_R0-0}{2}: [pubd] [30214]: (debug): Creating crimson data collector for filter '/process-cpu-ios-xe-oper:cpu-usage/cpu-utilization' (1 subfilters) with cap 0x0001.
2024/05/23 14:40:36.572671763 {pubd_R0-0}{2}: [pubd] [30214]: (debug): Need new data collector instance 0 for subfilter '/process-cpu-ios-xe-oper:cpu-usage/cpu-utilization'.
2024/05/23 14:40:36.572675434 {pubd_R0-0}{2}: [pubd] [30214]: (debug): Creating CRIMSON periodic data collector for filter '/process-cpu-ios-xe-oper:cpu-usage/cpu-utilization'.
2024/05/23 14:40:36.572688399 {pubd_R0-0}{2}: [pubd] [30214]: (debug): tree rooted at cpu-usage
2024/05/23 14:40:36.572715384 {pubd_R0-0}{2}: [pubd] [30214]: (debug): last container/list node 0
2024/05/23 14:40:36.572740734 {pubd_R0-0}{2}: [pubd] [30214]: (debug): 1 non leaf children to render from cpu-usage down
2024/05/23 14:40:36.573135594 {pubd_R0-0}{2}: [pubd] [30214]: (debug): URI:/cpu_usage;singleton_id=0 SINGLETON
2024/05/23 14:40:36.573147953 {pubd_R0-0}{2}: [pubd] [30214]: (debug): 0 non leaf children to render from cpu-utilization down
2024/05/23 14:40:36.573159482 {pubd_R0-0}{2}: [pubd] [30214]: (debug): Timer created for subscription 101, sensor 0x62551136f0e8
2024/05/23 14:40:36.573166451 {pubd_R0-0}{2}: [mdt-ctrl] [30214]: (note): {subscription receiver event='receiver connected'} received with peer (10.48.39.98:57000) for subscription 101 receiver 'grpc-tcp://10.48.39.98:57000'
2024/05/23 14:40:36.573197750 {pubd_R0-0}{2}: [pubd] [30214]: (debug): Starting batch from periodic collector 'ei_do periodic'.
2024/05/23 14:40:36.573198408 {pubd_R0-0}{2}: [pubd] [30214]: (debug): Building from the template
2024/05/23 14:40:36.575467870 {pubd_R0-0}{2}: [pubd] [30214]: (debug): Created dbal batch:133, for crimson subscription
2024/05/23 14:40:36.575470867 {pubd_R0-0}{2}: [pubd] [30214]: (debug): Done building from the template
2024/05/23 14:40:36.575481078 {pubd_R0-0}{2}: [pubd] [30214]: (debug): Executing batch:133 for periodic subscription
2024/05/23 14:40:36.575539723 {pubd_R0-0}{2}: [mdt-ctrl] [30214]: (note): {subscription id=101 receiver name='grpc-tcp://10.48.39.98:57000', state='connecting'} handling 'receiver connected' event with result 'e_mdt_rc_ok'
2024/05/23 14:40:36.575558274 {pubd_R0-0}{2}: [mdt-ctrl] [30214]: (note): {subscription receiver event='receiver connected'} subscription 101 receiver 'grpc-tcp://10.48.39.98:57000' changed
2024/05/23 14:40:36.577274757 {ndbmand_R0-0}{2}: [ndbmand] [30690]: (info): get__next_table reached the end of table for /services;serviceName=ewlc_oper/capwap_data@23
2024/05/23 14:40:36.577279206 {ndbmand_R0-0}{2}: [ndbmand] [30690]: (debug): Cleanup table for /services;serviceName=ewlc_oper/capwap_data cursor=0x57672da538b0
2024/05/23 14:40:36.577314397 {ndbmand_R0-0}{2}: [ndbmand] [30690]: (info): get__next_object cp=ewlc-oper-db exit return CONFD_OK
2024/05/23 14:40:36.577326609 {ndbmand_R0-0}{2}: [ndbmand] [30690]: (debug): yield ewlc-oper-db
2024/05/23 14:40:36.579099782 {iosrp_R0-0}{1}: [parser_cmd] [26295]: (note): id= A.B.C.D@vty0:user= cmd: 'receiver ip address 10.48.39.98 57000 protocol grpc-tcp' SUCCESS 2024/05/23 14:40:36.578 UTC
2024/05/23 14:40:36.580979429 {pubd_R0-0}{2}: [pubd] [30214]: (debug): Batch response received for crimson sensor, batch:133
2024/05/23 14:40:36.580988867 {pubd_R0-0}{2}: [pubd] [30214]: (debug): Green response: Result rc 0, Length: 360, num_records 1
2024/05/23 14:40:36.581175013 {pubd_R0-0}{2}: [pubd] [30214]: (debug): Green Resp cursor len 63
2024/05/23 14:40:36.581176173 {pubd_R0-0}{2}: [pubd] [30214]: (debug): There is no more data left to be retrieved from batch 133.
2024/05/23 14:40:36.581504331 {iosrp_R0-0}{2}: [parser_cmd] [24367]: (note): id= 10.227.65.133@vty1:user=admin cmd: 'receiver ip address 10.48.39.98 57000 protocol grpc-tcp' SUCCESS 2024/05/23 14:40:36.553 UTC
[...]
2024/05/23 14:40:37.173223406 {pubd_R0-0}{2}: [pubd] [30214]: (info): Added queue (wq: tc_inst 60293411, 101) to be monitored (mqid: 470)
2024/05/23 14:40:37.173226005 {pubd_R0-0}{2}: [pubd] [30214]: (debug): New subscription (subscription 101) monitoring object stored at id 19
2024/05/23 14:40:37.173226315 {pubd_R0-0}{2}: [pubd] [30214]: (note): Added subscription for monitoring (subscription 101, msid 19)
2024/05/23 14:40:37.173230769 {pubd_R0-0}{2}: [pubd] [30214]: (debug): Stats updated for Q (wq: tc_inst 60293411, 101), total_enqueue: 1
2024/05/23 14:40:37.173235969 {pubd_R0-0}{2}: [pubd] [30214]: (debug): (grpc::events) Processing event Q
2024/05/23 14:40:37.173241290 {pubd_R0-0}{2}: [pubd] [30214]: (debug): GRPC telemetry connector update data for subscription 101, period 1 (first: true)
2024/05/23 14:40:37.173257944 {pubd_R0-0}{2}: [pubd] [30214]: (debug): Encoding path is Cisco-IOS-XE-process-cpu-oper:cpu-usage/cpu-utilization
2024/05/23 14:40:37.173289128 {pubd_R0-0}{2}: [pubd] [30214]: (debug): Creating kvgpb encoder
2024/05/23 14:40:37.173307771 {pubd_R0-0}{2}: [pubd] [30214]: (debug): Creating combined parser
2024/05/23 14:40:37.173310050 {pubd_R0-0}{2}: [pubd] [30214]: (debug): Beginning MDT yang container walk for record 0
2024/05/23 14:40:37.173329761 {pubd_R0-0}{2}: [pubd] [30214]: (debug): Dispatching new container [data_node: name=Cisco-IOS-XE-process-cpu-oper:cpu-usage, type=container, parent=n/a, key=false]
2024/05/23 14:40:37.173334681 {pubd_R0-0}{2}: [pubd] [30214]: (debug): Container 'Cisco-IOS-XE-process-cpu-oper:cpu-usage' started successfully
2024/05/23 14:40:37.173340313 {pubd_R0-0}{2}: [pubd] [30214]: (debug): add data in progress
2024/05/23 14:40:37.173343079 {pubd_R0-0}{2}: [pubd] [30214]: (debug): GRPC telemetry connector continue data for subscription 101, period 1 (first: true)
2024/05/23 14:40:37.173345689 {pubd_R0-0}{2}: [pubd] [30214]: (debug): (grpc::events) Processing event Q
2024/05/23 14:40:37.173350431 {pubd_R0-0}{2}: [pubd] [30214]: (debug): Dispatching new container [data_node: name=cpu-utilization, type=container, parent=Cisco-IOS-XE-process-cpu-oper:cpu-usage, key=false]
2024/05/23 14:40:37.173353194 {pubd_R0-0}{2}: [pubd] [30214]: (debug): Deferred container cpu-utilization
2024/05/23 14:40:37.173355275 {pubd_R0-0}{2}: [pubd] [30214]: (debug): Container 'cpu-utilization' started successfully
2024/05/23 14:40:37.173380121 {pubd_R0-0}{2}: [pubd] [30214]: (debug): Dispatching new leaf [name=five-seconds, value=3, parent=cpu-utilization, key=false]
2024/05/23 14:40:37.173390655 {pubd_R0-0}{2}: [pubd] [30214]: (debug): Leaf 'five-seconds' added successfully
2024/05/23 14:40:37.173393529 {pubd_R0-0}{2}: [pubd] [30214]: (debug): add data in progress
2024/05/23 14:40:37.173395693 {pubd_R0-0}{2}: [pubd] [30214]: (debug): GRPC telemetry connector continue data for subscription 101, period 1 (first: true)
2024/05/23 14:40:37.173397974 {pubd_R0-0}{2}: [pubd] [30214]: (debug): (grpc::events) Processing event Q
2024/05/23 14:40:37.173406311 {pubd_R0-0}{2}: [pubd] [30214]: (debug): Dispatching new leaf [name=five-seconds-intr, value=0, parent=cpu-utilization, key=false]
2024/05/23 14:40:37.173408937 {pubd_R0-0}{2}: [pubd] [30214]: (debug): Leaf 'five-seconds-intr' added successfully
2024/05/23 14:40:37.173411575 {pubd_R0-0}{2}: [pubd] [30214]: (debug): add data in progress
[...]

確保指標到達TIG堆疊

從HonglumDB CLI

與其他資料庫系統一樣，ConfusionDB隨附一個CLI，可用於檢查Telegraf是否正確接收度量並儲存在定義的資料庫中。ConfusionDB將指標（稱為點）組織為度量，這些度量本身被組織為序列。此處介紹的一些基本命令可用於驗證ConsumeDB端的資料方案並確保資料到達此應用程式。

首先，您可以檢查序列、測量及其結構（鍵）是否正確產生。它們由Telegraf和ConsumeDB根據使用的RPC結構自動生成。

$ influx
Connected to http://localhost:8086 version 1.6.7~rc0
InfluxDB shell version: 1.6.7~rc0
> USE TELEGRAF
Using database TELEGRAF
> SHOW SERIES
key
---
Cisco-IOS-XE-process-cpu-oper:cpu-usage/cpu-utilization,host=ubuntu-virtual-machine,path=Cisco-IOS-XE-process-cpu-oper:cpu-usage/cpu-utilization,source=WLC,subscription=101
> SHOW MEASUREMENTS
name: measurements
name
----
Cisco-IOS-XE-process-cpu-oper:cpu-usage/cpu-utilization
> SHOW FIELD KEYS FROM "Cisco-IOS-XE-process-cpu-oper:cpu-usage/cpu-utilization"
name: Cisco-IOS-XE-process-cpu-oper:cpu-usage/cpu-utilization
fieldKey                                               fieldType
--------                                               ---------
cpu_usage_processes/cpu_usage_process/avg_run_time     integer
cpu_usage_processes/cpu_usage_process/five_minutes     float
cpu_usage_processes/cpu_usage_process/five_seconds     float
cpu_usage_processes/cpu_usage_process/invocation_count integer
cpu_usage_processes/cpu_usage_process/name             string
cpu_usage_processes/cpu_usage_process/one_minute       float
cpu_usage_processes/cpu_usage_process/pid              integer
cpu_usage_processes/cpu_usage_process/total_run_time   integer
cpu_usage_processes/cpu_usage_process/tty              integer
five_minutes                                           integer
five_seconds                                           integer
five_seconds_intr                                      integer
one_minute                                             integer

一旦資料結構明確（整數、字串、布林值、...），就可以根據特定欄位獲取在這些測量上儲存的資料點的數量。

# Get the number of points from "Cisco-IOS-XE-process-cpu-oper:cpu-usage/cpu-utilization" for the field "five_seconds".
> SELECT COUNT(five_seconds) FROM "Cisco-IOS-XE-process-cpu-oper:cpu-usage/cpu-utilization"
name: Cisco-IOS-XE-process-cpu-oper:cpu-usage/cpu-utilization
time count
---- -----
0    1170
> SELECT COUNT(five_seconds) FROM "Cisco-IOS-XE-process-cpu-oper:cpu-usage/cpu-utilization"
name: Cisco-IOS-XE-process-cpu-oper:cpu-usage/cpu-utilization
time count
---- -----
0    1171

# Fix timestamp display
> precision rfc3339
# Get the last point stored in "Cisco-IOS-XE-process-cpu-oper:cpu-usage/cpu-utilization" for the field "five_seconds".
> SELECT LAST(five_seconds) FROM "Cisco-IOS-XE-process-cpu-oper:cpu-usage/cpu-utilization"
name: Cisco-IOS-XE-process-cpu-oper:cpu-usage/cpu-utilization
time                    last
----                    ----
2024-05-23T13:18:53.51Z 0
> SELECT LAST(five_seconds) FROM "Cisco-IOS-XE-process-cpu-oper:cpu-usage/cpu-utilization"
name: Cisco-IOS-XE-process-cpu-oper:cpu-usage/cpu-utilization
time                     last
----                     ----
2024-05-23T13:19:03.589Z 2

如果特定欄位的點數和上次出現的時間戳增加，則TIG堆疊正確接收和儲存WLC傳送的資料是很好的跡象。

從Telegraf

要驗證Telegraf接收方確實從控制器獲取了一些度量並檢查其格式，您可以將Telegraf度量重定向到主機上的輸出檔案。在裝置互連故障排除方面，這非常方便。為此，只需使用Telegraf的「file」輸出外掛(可透過/etc/telegraf/telegraf.conf進行配置)。

# Send telegraf metrics to file(s)
[[outputs.file]]
#   ## Files to write to, "stdout" is a specially handled file.
   files = ["stdout", "/tmp/metrics.out", "other/path/to/the/file"]
#
#   ## Use batch serialization format instead of line based delimiting.  The
#   ## batch format allows for the production of non line based output formats and
#   ## may more efficiently encode metric groups.
#   # use_batch_format = false
#
#   ## The file will be rotated after the time interval specified.  When set
#   ## to 0 no time based rotation is performed.
#   # rotation_interval = "0d"
#
#   ## The logfile will be rotated when it becomes larger than the specified
#   ## size.  When set to 0 no size based rotation is performed.
#   # rotation_max_size = "0MB"
#
#   ## Maximum number of rotated archives to keep, any older logs are deleted.
#   ## If set to -1, no archives are removed.
#   # rotation_max_archives = 5
#
#   ## Data format to output.
#   ## Each data format has its own unique set of configuration options, read
#   ## more about them here:
#   ## https://github.com/influxdata/telegraf/blob/master/docs/DATA_FORMATS_OUTPUT.md
   data_format = "influx"

參考資料

硬體大小調整準則

格拉法納要求