I-automate amazon eks Ukuxazulula inkinga usebenzisa ukuhamba komsebenzi we-Amazon Bedrock Agent

Njengoba izinhlangano zikala izinsiza zabo ze-Amazon Enastic Kubernetes (Ama-Amazon Eks), abaphathi bama-Platform babhekene nezinselelo ezikhulayo ekuphatheni kahle amaqembu ahlukahlukene. Imisebenzi enjengokuphenya ukwehluleka kwe-pod, ukubhekana nezinkinga zensiza, nokuxazulula ukungasebenzi kahle kungadla isikhathi nomzamo obalulekile. Esikhundleni sokusebenzisa amahora abalulekile wobunjiniyela onobunjiniyela ngokwakho izingodo, ukulandelela amamethrikhi, kanye nokusebenzisa ukulungiswa, amaqembu kufanele agxile ekushayeleni okusha. Manje, ngamandla e-Advocate AI, ungaguqula imisebenzi yakho ye-Kubernetes. Ngokusebenzisa ukuqapha okuhlakaniphile kweqembu, ukuhlaziya iphethini, kanye nokulungiswa okuzenzakalelayo, unganciphisa kakhulu zombili izikhathi ezisho ukuthi ukukhomba (MTTTI) kanye nesikhathi sokuxazulula (MTTR) ngezinkinga ezijwayelekile zeqoqo.

Ku-AWS Re: Ukusungula i-2024, samemezela amandla okubambisana e-multi-alent for Amazon Bedrock (ukubuka kuqala). Ngokusebenzisana ngama-alent multi, ungakha, ukuhambisa, nokuphatha ama-agents amaningi e-AI asebenza ndawonye emisebenzini eyinkimbinkimbi eminingi edinga amakhono akhethekile. Ngoba ukuxazulula inkinga Iqoqo le-EKS libandakanya ukutholwa kokuthola ulwazi kusuka kuma-Pipeline amaningi okubukwa kanye nokusebenzisa amapayipi aqhubekayo nokuthunyelwa (CI / CD) Ipayipi, ukugeleza komsebenzi okunemisebenzi eminingi kungasiza iqembu le-Operations liqondise ukuphathwa kwamaqembu e-EKS. Umphathi weMenenja Womsebenzi ungahlanganisa nama-ejenti ngamanye abonisa isikhombimsebenzisi sokubukwa komuntu ngamunye kanye nokuhamba komsebenzi we-CI / CD ukuhlela ama-orchestrate nokwenza imisebenzi esekelwe ekushesheni komsebenzisi.

Kulokhu okuthunyelwe, sibonisa ukuthi ungahlela kanjani ama-Amazon Bedrock Agents amaningi ukudala uhlelo oluyinkimbinkimbi ama-Amazon EKS Xining system. Ngokunika amandla ukusebenzisana phakathi kwemininingwane ekhethekile etholakele kusuka ku-K8Sgpt kanye nokwenza izinto ngohlaka lwe-Argocd-Ungakha okuguquguqukayo okubonakalayo okukhomba, ukuhlaziya, nokuxazulula futhi kuxazululwe izingqinamba ze-Cluster ngokungenelela komuntu okuncane.

Ukubuka konke

Ukwakhiwa kuqukethe izinto ezilandelayo eziyisisekelo:

• I-Amazon Bedrock Collisitor Agent – i-Orchestrates Ukuhamba Kwomsebenzi futhi ugcina umongo ngenkathi umshukumisa umsebenzisi ushukumisela kuma-ejenti akhethekile, ukuphatha ukusebenza kwe-multistep kanye nokusebenzisana kwe-ejenti
• I-Amazon Bedrock Agent ye-K8SGTT – Ihlola imicimbi ye-Cluster ne-POD ngokusebenzisa i-K8SGPT's Acque API ngezinkinga zokuphepha, ama-Admications, kanye nezinkinga zokusebenza, ukuhlinzeka ngeziphakamiso zokulungisa ngolimi lwemvelo
• I-Agent ye-Amazon Bedrock ye-AROSCD – ilawula ukulungiswa okususelwa ku-Gitops nge-Argocd, ukuphatha ama-rollbacks, insizakalo yezinsizakusebenza, nokuvuselelwa kokucushwa

Umdwebo olandelayo ukhombisa ukwakhiwa kwesixazululo.

Izimfuneko

Udinga ukuba nezimfuneko ezilandelayo endaweni elandelayo:

Setha i-Amazon Eks Cluster nge-K8Sgpt kanye ne-Argocs

Siqala ngokufaka nokulungiselela i-K8Sgpt opharetha kanye nesilawuli se-Argocd ku-Eks Cluster.

Umphathi we-K8Sgpt uzosiza ngokunika amandla ukuhlaziya amandla we-AI nokuxazulula inkinga yezinkinga zeqoqo. Isibonelo, ingabona ngokuzenzakalelayo futhi kusikisele ukulungiswa kokuthunyelwa okungalungile, njengokuhlonza nokuxazulula izinkinga zokucindezela izinsiza kuma-pods.

I-Argocd iyithuluzi elimemezela i-Gitops Ithuluzi eliqhubekayo lokulethwa kwama-kubernetetes athisha ukuthunyelwa kwezicelo ngokugcina isimo sohlelo lokusebenza olufunayo ekuvumelaniswe nalokho okuchazwe endaweni yokugcina izinto.

I-Agent ye-Amazon Bedrock isebenza njengoMentu ohlakaniphile wokuthatha izinqumo ekwakhiweni kwethu, ekuhlaziyeni izingqinamba zeqoqo ezitholwe yi-K8SGTT. Ngemuva kwembangela yezimpande kukhonjwa, izenzo zokulungisa i-ejenti ye-ejenti nge-argocd's Gitops Injini. Lokhu kuhlanganiswa okunamandla kusho ukuthi lapho kutholwa izinkinga (noma ngabe ukuthunyelwa okungalungile, izingqinamba zezinsiza, noma inkinga yokulinganisa), i-ejenti ingahlanganiswa ngokuzenzakalela ne-Arrocd ukunikeza ukulungisa okudingekayo. I-Argocd Bese ithatha lezi zinguquko futhi ivumelanise neqembu lakho le-EKS, kwakha ingqalasizinda yokuzelapha ngempela.

1. Dala izindawo ezidingekayo kuma-Amazon EKS:

   kubectl create ns helm-guestbook
   kubectl create ns k8sgpt-operator-system

2. Faka i-k8sgpt helm repository bese ufaka opharetha:

   helm repo add k8sgpt 
   helm repo update
   helm install k8sgpt-operator k8sgpt/k8sgpt-operator 
     --namespace k8sgpt-operator-system

3. Ungaqinisekisa ukufakwa ngokufaka umyalo olandelayo:

   kubectl get pods -n k8sgpt-operator-system
   
   NAME                                                          READY   STATUS    RESTARTS  AGE
   release-k8sgpt-operator-controller-manager-5b749ffd7f-7sgnd   2/2     Running   0         1d
   

Ngemuva kokuthi opharetha bathunyelwe, ungahlela insiza ye-K8SGTT. Le ncazelo yangokwezifiso yezezimali (i-CRD) izoba nokulungiswa okukhulu kwezilimi (LLM) ezosiza ekuhlaziyweni okunamandla kwe-AI nokuxazulula inkinga yezinkinga zeqoqo. I-K8Sgpt isekela izipele ezahlukahlukene ukusiza ekuhlaziyweni okunamandla kwe-AI. Kulokhu okuthunyelwe, sisebenzisa i-Amazon Bedrock njenge-Backend kanye ne-Anthropic's Claude V3 njenge-LLM.

4. Udinga ukudala ubunikazi be-pod ngokuhlinzeka ukufinyelela kwama-eks cluster kwamanye ama-AWS Services nge-Amazon Bedrock:

   eksctl create podidentityassociation  --cluster PetSite --namespace k8sgpt-operator-system --service-account-name k8sgpt  --role-name k8sgpt-app-eks-pod-identity-role --permission-policy-arns arn:aws:iam::aws:policy/AmazonBedrockFullAccess  --region $AWS_REGION

5. Lungiselela i-K8SGPT CRD:

   cat << EOF > k8sgpt.yaml
   apiVersion: core.k8sgpt.ai/v1alpha1
   kind: K8sGPT
   metadata:
     name: k8sgpt-bedrock
     namespace: k8sgpt-operator-system
   spec:
     ai:
       enabled: true
       model: anthropic.claude-v3
       backend: amazonbedrock
       region: us-east-1
       credentials:
         secretRef:
           name: k8sgpt-secret
           namespace: k8sgpt-operator-system
     noCache: false
     repository: ghcr.io/k8sgpt-ai/k8sgpt
     version: v0.3.48
   EOF
   
   kubectl apply -f k8sgpt.yaml
   

6. Qinisekisa izilungiselelo ukuqinisekisa i-K8SGPT-Bedrock Pod pod isebenza ngempumelelo:

   kubectl get pods -n k8sgpt-operator-system
   NAME                                                          READY   STATUS    RESTARTS      AGE
   k8sgpt-bedrock-5b655cbb9b-sn897                               1/1     Running   9 (22d ago)   22d
   release-k8sgpt-operator-controller-manager-5b749ffd7f-7sgnd   2/2     Running   3 (10h ago)   22d
   

7. Manje ungahlela isilawuli se-Argucd:

   helm repo add argo 
   helm repo update
   kubectl create namespace argocd
   helm install argocd argo/argo-cd 
     --namespace argocd 
     --create-namespace

8. Qinisekisa ukufakwa kwe-Arpocd:

   kubectl get pods -n argocd
   NAME                                                READY   STATUS    RESTARTS   AGE
   argocd-application-controller-0                     1/1     Running   0          43d
   argocd-applicationset-controller-5c787df94f-7jpvp   1/1     Running   0          43d
   argocd-dex-server-55d5769f46-58dwx                  1/1     Running   0          43d
   argocd-notifications-controller-7ccbd7fb6-9pptz     1/1     Running   0          43d
   argocd-redis-587d59bbc-rndkp                        1/1     Running   0          43d
   argocd-repo-server-76f6c7686b-rhjkg                 1/1     Running   0          43d
   argocd-server-64fcc786c-bd2t8                       1/1     Running   0          43d

9. Patch insizakalo ye-Argecd ukuze ube ne-Barternal Load Balancer:

   kubectl patch svc argocd-server -n argocd -p '{"spec": {"type": "LoadBalancer"}}'

10. Manje usungafinyelela i-UI ye-Argecd nge-Load Balancer Endpoint kanye neziqinisekiso zomsebenzisi wokuphatha:

    kubectl get svc argocd-server -n argocd
    NAME            TYPE           CLUSTER-IP       EXTERNAL-IP                                                              PORT(S)                      AGE
    argocd-server   LoadBalancer   10.100.168.229   a91a6fd4292ed420d92a1a5c748f43bc-653186012.us-east-1.elb.amazonaws.com   80:32334/TCP,443:32261/TCP   43d

11. Phinda uthole ubufakazi be-UI ye-Arspocd:

    export argocdpassword=`kubectl -n argocd get secret argocd-initial-admin-secret 
    -o jsonpath="{.data.password}" | base64 -d`
    
    echo ArgoCD admin password - $argocdpassword

12. Cindezela ubuqiniso ku-AWS Searte Manager:

    aws secretsmanager create-secret 
    --name argocdcreds 
    --description "Credentials for argocd" 
    --secret-string "{"USERNAME":"admin","PASSWORD":"$argocdpassword"}"

13. Lungiselela uhlelo lokusebenza lwesampula e-Agocd:

    cat << EOF > argocd-application.yaml
    apiVersion: argoproj.io/v1alpha1
    kind: Application
    metadata:
    name: helm-guestbook
    namespace: argocd
    spec:
    project: default
    source:
    repoURL: 
    targetRevision: HEAD
    path: helm-guestbook
    destination:
    server: 
    namespace: helm-guestbook
    syncPolicy:
    automated:
    prune: true
    selfHeal: true
    EOF

14. Faka isicelo sokucushwa futhi uqinisekise kusuka ku-UI Argucd UI ngokungena ngemvume njengomsebenzisi wokuphatha:

    kubectl apply -f argocd-application.yaml

    

15. Kuthatha isikhathi se-ksggpt ukuhlaziya ama-pods asanda kudalwa. Ukwenza lokho ngokushesha, qala kabusha ama-pods adalwe kwi-K8SGPT-Operator-System Systempace. Ama-pods angaqalwa kabusha ngokufaka umyalo olandelayo:

    kubectl -n k8sgpt-operator-system rollout restart deploy
    
    deployment.apps/k8sgpt-bedrock restarted
    deployment.apps/k8sgpt-operator-controller-manager restarted

Setha ama-ejenti we-Amazon Bedrock Agents we-K8Sgpt kanye ne-Argocs

Sisebenzisa isitaki se-Cloudformation ukufaka ama-ejenti ngamanye eSifundeni sase-US East (N. Virginia). Lapho ufaka ithempulethi ye-Cloudformation, uthumela izinsiza eziningana (izindleko zizotholwa ngezinsizakusebenza ze-AWS ezisetshenzisiwe).

Sebenzisa amapharamitha alandelayo wethempulethi ye-Cloudformation:

Isitaki sakha imisebenzi elandelayo ye-AWS Lambda:

• -LambdaK8sGPTAgent-
• -RestartRollBackApplicationArgoCD-
• -ArgocdIncreaseMemory-

Isitaki sakha ama-ejenti alandelayo e-Amazon Bedrock Agents:

• ArgoCDAgentngamaqembu ezenzweni alandelayo:
  1. argocd-rollback
  2. argocd-restart
  3. argocd-memory-management

• K8sGPTAgentneqembu elilandelayo lesenzo:
  1. k8s-cluster-operations

Isitaki sikhipha okulandelayo, nalawa manzelelo alandelayo ahambisana nalo:

1. ArgoCDAgent
2. K8sGPTAgent

• I-Lambdak8Sggpptagentrole, ubunikazi be-AWS ubunikazi kanye nokuphathwa kokufinyelela (i-IAM) Iphakethe le-Amazon Resourvation Igama (ARN) elihlotshaniswa nomsebenzi weLambda ophethe ukusebenzisana ne-K8SGPT Agent ku-ES Cluster. Le ndima i-ARN izodingeka esigabeni sakamuva senqubo yokucushwa.
• K8sGPTAgentAliasIdI-ID ye-K8SGPT IZOMEDROCK ALER ALERT ALIAS ALIAS
• ArgoCDAgentAliasIdI-ID ye-Assold Amazon Bedrock Agent Alias ​​Alias
• CollaboratorAgentAliasIdI-ID ye-Agent Agent Alias

Nikeza izimvume ezifanele ukunika amandla i-K8Sgpt Amazon Bedrock Agent ukufinyelela i-EKS Cluster

Ukuze unike amandla i-K8SGPTICTION MEDROCK Agent ukufinyelela i-ejenti ye-EKS, udinga ukumisa izimvume ze-IAM ezifanele usebenzisa i-Amazon Eks Excice Management Management API. Le yinqubo yezinyathelo ezimbili: Okokuqala, udala ukungena kokufinyelela kweqhaza lomsebenzi we-Lambda (ongakuthola esigabeni sokukhipha isifanekiso), bese uhlobanisa AmazonEKSViewPolicy ukunikeza ukufinyelela okufundwayo kuphela kwiqoqo. Lokhu kucushwa kuqinisekisa ukuthi i-ejenti ye-K8SGPT inezimvume ezidingekayo zokuqapha kanye nokuhlaziya izinsizakusebenza ze-EKS Cluster ngenkathi zigcina umgomo wokuthola ilungelo elincane.

1. Dala ukungena kokufinyelela kweqhaza le-Lambda Umsebenzi we-Lambda

   export CFN_STACK_NAME=EKS-Troubleshooter
   	   export EKS_CLUSTER=PetSite
   
   export K8SGPT_LAMBDA_ROLE=`aws cloudformation describe-stacks --stack-name $CFN_STACK_NAME --query "Stacks[0].Outputs[?OutputKey=='LambdaK8sGPTAgentRole'].OutputValue" --output text`
   
   aws eks create-access-entry 
       --cluster-name $EKS_CLUSTER 
       --principal-arn $K8SGPT_LAMBDA_ROLE

2. Hlanganisa inqubomgomo yokubuka e-EKS nokungena kokufinyelela

   aws eks associate-access-policy 
       --cluster-name $EKS_CLUSTER 
       --principal-arn  $K8SGPT_LAMBDA_ROLE
       --policy-arn arn:aws:eks::aws:cluster-access-policy/AmazonEKSClusterAdminPolicy 
       --access-scope type=cluster

3. Qinisekisa ama-Amazon Bedrock Agents. Isifanekiso se-Cloudformation sengeza wonke ama-ejenti amathathu adingekayo. Ukubuka ama-ejenti, e-Amazon Bedrock Console, ngaphansi kwamathuluzi omakhi kwifasigi yokuhambisa, khetha Ama-ejentinjengoba kukhonjisiwe ku-skrini elandelayo.

Yenza i-Amazon ECK Ukuxazulula inkinga usebenzisa ukuhamba komsebenzi we-Amazon Bedrock

Manje, vivinya ikhambi. Sihlola lezi zimo ezimbili ezilandelayo:

1. I-ejenti ixhumanisa ne-K8SGPT ASTENT ukuhlinzeka ngokuqonda esimweni sezimpande sokwehluleka kwe-pod
2. I-Agent Collaborator ixhumanisa ne-ejenti ye-Arpocd ukuhlinzeka ngempendulo

I-Agent ixhumanisa ne-K8SGPT ASEN NETENT ukuhlinzeka ngokuqonda esimweni sezimpande sokwehluleka kwe-pod

Kulesi sigaba, sihlola ukuqaphela phansi kwesicelo sesampula esibizwa ngeMemory-Demo. Sinentshisekelo kwimbangela yenkinga. Sisebenzisa lokhu okulandelayo: “Sithole ukuqaphela phansi kohlelo lokusebenza lweMemory-Demo. Sisize ngembangela yenkinga.”

Umenzeli akagcinanga nje kuphela imbangela, kodwa wahamba ngesinyathelo esisodwa ukuze alungise iphutha, okulesi simo ukwandisa izinsizakusebenza zenkumbulo kuhlelo lokusebenza.

I-Agent esebenzisana nayo ixhumanisa ne-ejenti ye-Argocd ukuhlinzeka ngempendulo

Ngalesi simo, siyaqhubeka nokusuka phambili. Sibona sengathi isicelo asinikezwanga inkumbulo eyanele, futhi kufanele inyuswe ukuze ilungise inkinga unomphela. Singasho futhi uhlelo lokusebenza lusesimweni esingenampilo e-UI Argocd UI, njengoba kukhonjisiwe ku-skrini elandelayo.

Manje ake siqhubeke ukukhulisa inkumbulo, njengoba kukhonjisiwe ku-skrini elandelayo.

I-ejenti exhumane ne argocd_operations I-Agent ye-Amazon Bedrock futhi yakwazi ukukhulisa ngempumelelo inkumbulo. Kufana okufanayo kungahle kuthathwe e-UI ye-Arspocd UI.

Hlanza

Uma uthatha isinqumo sokuyeka ukusebenzisa ikhambi, qedela lezi zinyathelo ezilandelayo:

1. Ukususa izinsiza ezihambisanayo ezisetshenziswe kusetshenziswa i-AWS Cloudformation:
   1. Ku-AWS Cloud Cloudform Console, khetha izitaki kufasitelana navigation.
   2. Thola isitaki osidalile ngesikhathi senqubo yokuhambisa (owabelwe igama).
   3. Khetha isitaki bese ukhetha Susa.
2. Susa i-eks cluster uma udale eyodwa ngokukhethekile lokhu kuqaliswa.

Ukugcina

Ngokufaka ama-ejenti amaningi we-Amazon Bedrock Agents, sikhombise ukuthi singawakha kanjani uhlelo lokuxazulula inkinga ye-AI-Powered Equament EMS lukwenza lula ukusebenza kwe-Kubernetes. Lokhu kuhlanganiswa kokuhlaziywa kwe-k8sgpt kanye nokuzenzakalela okuzenzakalelayo kukhombisa amathuba anamandla lapho kuhlanganisa abenzeli abakhethekile be-AI ngamathuluzi akhona we-devops. Yize lesi sixazululo simele ukuthuthuka emisebenzini ezenzakalelayo ku-Kubernetetes, kubalulekile ukukhumbula ukuthi ukwengamela komuntu kuhlala kukubalulekile, ikakhulukazi izimo eziyinkimbinkimbi nezinqumo ezinhle.

Njengoba i-Amazon Bedrock kanye namakhono ayo e-ejenti ayaqhubeka nokuvela, singalindela amathuba ama-orchestration amaningi kakhulu. Ungangezelela lesi sixazululo ukufaka amathuluzi angeziwe, amamethrikhi, kanye nokuhamba komsebenzi okuzenzakalelayo ukuze uhlangabezane nezidingo ezithile zenhlangano yakho.

Ukuze ufunde kabanzi nge-Amazon Bedrock, bheka izinsiza ezilandelayo:

Mayelana nababhali

Vikram venkataraman Ingabe ukwakhiwa kwezixazululo eziphikisanayo ezikhethekile ezinsizakalweni zewebhu zase-Amazon (AWS). Usiza amakhasimende ngokushesha, ukukala, futhi amukele imikhuba emihle yemisebenzi yabo efakiwe. Ngokuvela kwe-ani ekhiqizayo, i-vikram isebenze ngenkuthalo namakhasimende ukuze aqinisekise izinsizakalo ze-AWS / ML ze-AWS ukuxazulula izinselelo eziyinkimbinkimbi zokusebenza, ukuguquguquka kokushintshana kokusebenza, kanye nokwenza ngcono ukuphendula kwezigameko nge-automation ehlakaniphile.

UPuneeth Ranjan Komaragiri ngumphathi we-akhawunti oyinhloko wezobuchwepheshe e-Amazon Web Services (AWS). Unothando olukhulu ngokuqapha nokuqashelwa, ukuphathwa kwezimali, kanye nezizinda ze-AI ezikhiqizayo. Endimeni yakhe yamanje, uPuneiseth uthokozela ukusebenzisana eduze namakhasimende, athambe ubuchwepheshe bakhe ukuze abasize baklame futhi bakhiphe imithwalo yabo yemisebenzi yabo yefu ngezinga elifanele nokuqina.

Sudheer sangunni ngumphathi we-akhawunti ephezulu yezobuchwepheshe e-AWS Enterprise ukwesekwa. Ngobuchwepheshe bakhe obukhulu efwini le-AWS kanye nedatha enkulu, iSudheer idlala indima ebaluleke kakhulu ekusizeni amakhasimende ngokuthuthukisa amandla abo okuqapha kanye nokubukwa kokunikezwa ngaphakathi kweminikelo ye-AWS.

IVikrant Choudhary ngumphathi we-akhawunti ophezulu wezobuchwepheshe e-Amazon Web Services (AWS), ochwepheshe kwezempilo kanye nesayensi yempilo. Ngaphezulu kweminyaka engu-15 yesipiliyoni kwizixazululo zamafu kanye nezakhiwo zebhizinisi, usiza amabhizinisi asheshise izindlela zawo zokuguqula izidakamizwa. Endimeni yakhe yamanje, ozakwethu bakwaVikrant namakhasimende azokwakha kanye nokusebenzisa izisombululo ezintsha, kusukela kumafu asungula ubuchwepheshe obusafufusa ai, ukushayela imiphumela yebhizinisi ephumelelayo.