Ikusasa lama-AIOps: Ukuhlanganisa i-AI yokuphatha okusebenzayo kwe-IT | B4N1

Ikusasa lama-AIOps: Ukuhlanganisa i-AI yokuphatha okusebenzayo kwe-IT

9 months ago · b4n1 Read time: 3 min

Isingeniso

Ukuthuthuka okusheshayo kobuchwepheshe kuholele ekukhuleni okubonakalayo kubunkimbinkimbi bezinhlelo zesimanje ze-IT. Ngenxa yalokho, amaqembu e-IT abhekene nezinselele ezingakaze zibonwe ekuphatheni nasekugcineni lezi zinhlelo, okuholela ekwenyukeni kwesikhathi sokunciphisa, ukwehla kokukhiqiza, kanye nezindleko eziphezulu zokusebenza. Ukuze kubhekwane nalezi zinselele, umqondo we-Artificial Intelligence for IT Operations (AIOps) uqhamuke njengoshintsho lwegeyimu. I-AIOps ithuthukisa ukufunda komshini kanye ne-AI ukuze iguqule ukugeleza komsebenzi we-DevOps, inike amandla ukuphathwa kwe-IT okusebenzayo, ukuhlaziya okubikezelayo, kanye ne-automation.

Kulesi sihloko, sizohlola ikusasa le-AIOps, izinzuzo zayo, namasu okusebenzisa. Sizophinde sihlole ocwaningweni lwangempela lomhlaba futhi sinikeze ibhulokhi yekhodi eyinkimbinkimbi ukuze sibonise ukusetshenziswa kwama-AIOps ekuphathweni kwe-IT.

Yini i-AIOps?

I-AIOps iyinkambu evelayo ehlanganisa amandla obuhlakani bokwenziwa (AI) nokufunda komshini (ML) nokusebenza kwe-IT (ITOps) ukuze kuthuthukiswe ukusebenza kahle, ukusebenza kahle, nokuthembeka kwamasevisi e-IT. Izinkundla ze-AIOps zisebenzisa ukuhlaziya idatha, ukumodela okubikezelwayo, kanye ne-automation ukuhlonza nokuxazulula izinkinga ze-IT ngaphambi kokuba zibe izinkinga ezinkulu.

Ama-AIOps angasetshenziswa ezindaweni ezahlukahlukene zokuphatha i-IT, okubandakanya:

  1. Izibalo Eziqagelayo: Izinkundla ze-AIOps zisebenzisa ama-algorithms okufunda komshini ukuze zihlaziye idatha yomlando futhi zibikezele izinkinga ezingaba khona ze-IT, zinike amandla ukugcinwa okumatasa nokunciphisa isikhathi sokuphumula.
  2. Ukulawulwa Kwesigameko: I-AIOps ingathola futhi ixazulule ngokuzenzekelayo izehlakalo ze-IT, inciphise isikhathi esimaphakathi sokuhlonza (MTTD) kanye nesikhathi esisho ukuxazulula (MTTR).
  3. Ukuzenzakalela: I-AIOps ingakwazi ukwenza imisebenzi evamile ngokuzenzakalelayo, njengokuphatha ukuhlela, ukuphathwa kwezichibi, nokwenza ikhophi yasenqolobaneni nokuthola usizo.

Izinzuzo ze-AIOps

Izinzuzo ze-AIOps ziningi futhi zibalulekile:

  1. Ukusebenza Okuthuthukisiwe: I-AIOps izenzela ngokuzenzakalelayo imisebenzi evamile, ikhulula abasebenzi be-IT ukuze bagxile ezinhlelweni zamasu.
  2. Ukuthembeka Okuthuthukisiwe: I-AIOps ibikezela futhi ivimbele izinkinga ze-IT, yehlisa isikhathi sokuphumula futhi ithuthukisa ukwethembeka okuphelele kwesistimu.
  3. Ukukhiqiza Okukhulisiwe: I-AIOps inika amandla amaqembu e-IT ukuthi aphendule ngokushesha ezehlakalweni, anciphise isikhathi esimaphakathi sokuhlonza (MTTD) kanye nesikhathi esisho ukuxazulula (MTTR).
  4. Ukwenziwa Kwezinqumo Okungcono: I-AIOps ihlinzeka ngamathimba e-IT ngemininingwane yesikhathi sangempela kanye nezibalo eziqagelayo, okuvumela ukwenza izinqumo okuqhutshwa idatha.

Amasu Okusebenzisa

Ukusebenzisa ama-AIOps kudinga indlela yesu:

  1. Ukuqoqwa Kwedatha: Qoqa idatha ye-IT emithonjeni ehlukahlukene, okuhlanganisa amalogi, amamethrikhi, namafayela okulungiselela.
  2. Ukudidiyelwa Kwedatha: Hlanganisa idatha esuka emithonjeni ehlukahlukene iye endaweni eyodwa.
  3. Ukufunda Ngomshini: Sebenzisa ama-algorithms okufunda komshini ukuze uhlaziye idatha futhi ubikezele izinkinga ezingaba khona ze-IT.
  4. Ukuzenzakalela: Yenza imisebenzi evamile kanye nokugeleza komsebenzi ngokuzenzakalelayo.
  5. Ukuqapha Nempendulo: Qhubeka uqaphe ukusebenza kwe-AIOps futhi unikeze impendulo ukuze uthuthukise ukunemba nokusebenza kahle.

Izifundo ze-Real-World

Izinhlangano eziningi zenze ngempumelelo ama-AIOps ukuguqula ukuhamba komsebenzi wokuphatha kwe-IT:

  1. I-Verizon: I-Verizon isebenzise i-AIOps ukuze ithuthukise ukwethembeka kwenethiwekhi nokunciphisa isikhathi sokuphumula. Inkundla isebenzise ukufundwa komshini ukuze ibikezele futhi ivimbele izinkinga zenethiwekhi, okuholele ekuncishisweni kwesikhathi esingu-30%.
  2. AT&T: I-AT&T isebenzise i-AIOps ukuze ithuthukise ukuphathwa kwesigameko futhi yehlise isikhathi esimaphakathi sokutholwa (i-MTTD) kanye nesikhathi esimaphakathi sokuxazulula (MTTR). Isiteji sasebenzisa i-automation ukuze ithole futhi ixazulule izehlakalo ze-IT, okuholele ekunciphiseni kwe-MTTD ngo-40% kanye nokwehliswa kwe-50% ku-MTTR.
  3. I-Microsoft: I-Microsoft isebenzise i-AIOps ukuze ithuthukise ukuhlaziya okubikezelayo kanye nokuzenzakalela. Inkundla isebenzise ukufundwa komshini ukubikezela izinkinga ezingaba khona ze-IT, inike amandla ukunakekelwa okusebenzayo kanye nokunciphisa isikhathi sokuphumula.

Vimba Ikhodi: Isibonelo se-AIOps kuPython

Nali ibhulokhi yekhodi eyinkimbinkimbi kuPython ekhombisa ukusetshenziswa kwe-AIOps ekuphathweni kwe-IT:

ngenisa ama-panda njenge-pd
kusuka ku-sklearn.ensemble ngenisa i-RandomForestClassifier
kusuka ku-sklearn.model_selection yokungenisa train_test_split
kusuka ku-sklearn.metrics import accuracy_score

# Layisha idatha
idatha = pd.read_csv('it_data.csv')

#Lungisa kusengaphambili idatha
X = idatha.drop(['isigameko'], i-eksisi=1)
y = idatha['isigameko']

# Hlukanisa idatha kumasethi okuqeqesha nawokuhlola
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

# Qeqesha i-classifier yehlathi engahleliwe
rfc = RandomForestClassifier(n_estimators=100, random_state=42)
rfc.fit(X_train, y_train)

# Yenza izibikezelo
y_pred = rfc.predict(X_test)

# Linganisa imodeli
ukunemba = accuracy_score(y_test, y_pred)
phrinta(f'Ukunemba: {ukunemba:.2f}')

# Sebenzisa imodeli ukubikezela izinkinga ezingaba khona ze-IT
new_data = pd.DataFrame({'cpu_usage': [80], 'memory_usage': [90]})
isibikezelo = rfc.predict(new_data)
uma isibikezelo[0] == 1:
    phrinta('Inkinga ye-IT engaba khona itholiwe!')

Lesi sakhiwo sekhodi sisebenzisa ukuhlelwa kwehlathi okungahleliwe ukubikezela izinkinga ezingaba khona ze-IT ngokusekelwe ku-CPU nedatha yokusetshenziswa kwememori. Imodeli iqeqeshwe kudatha yomlando futhi isetshenziselwe ukwenza izibikezelo kudatha entsha.

Ukubona ngeso lengqondo: I-AIOps Platform Architecture

Nakhu okubonwayo kwesakhiwo seplathifomu ye-AIOps kusetshenziswa i-Chart.js:

{
  "type": "pieChart",
  "idatha": {
    "labels": ["Ukuqoqwa Kwedatha", "Ukuhlanganiswa Kwedatha", "Ukufunda Ngomshini", "Ukuzenzakalela", "Ukuqapha kanye Nempendulo"],
    "idathasethi": [
      {
        "label": "AIOps Platform Architecture",
        "idatha": [30, 20, 20, 15, 15],
        "backgroundColor": ["#FF6384", "#36A2EB", "#FFCE56", "#4BC0C0", "#FF99CC"],
        "hoverBackgroundColor": ["#FF99CC", "#36A2EB", "#FFCE56", "#4BC0C0", "#FF99CC"]
      }
    ]
  },
  "izinketho": {
    "isihloko": {
      "Display": iqiniso,
      "text": "I-AIOps Platform Architecture"
    },
    "i-legend": {
      "isibonisi": iqiniso
    }
  }
}

Lokhu kuboniswa kubonisa izingxenye ezihlukene zeplathifomu ye-AIOps yezakhiwo, okuhlanganisa ukuqoqwa kwedatha, ukuhlanganiswa kwedatha, ukufunda ngomshini, ukuzenzakalela, nokuqapha kanye nempendulo.

Isiphetho

I-AIOps ubuchwepheshe obushintsha umdlalo obushintsha indlela amaqembu e-IT aphatha ngayo futhi agcine amasistimu e-IT ayinkimbinkimbi. Ngokusebenzisa ukufunda komshini kanye ne-AI, i-AIOps inika amandla ukuphathwa kwe-IT okusebenzayo, ukuhlaziya okubikezelayo, nokuzenzakalela, okuholela ekusebenzeni okuthuthukisiwe, ukuthembeka okuthuthukisiwe, kanye nokwanda kokukhiqiza. Ngokusebenzisa ama-AIOps, izinhlangano zinganciphisa isikhathi sokuphumula, zithuthukise ukwenza izinqumo, futhi zithole umkhawulo wokuncintisana emakethe.

Njengoba ubuchwepheshe buqhubeka nokuvela, singalindela ukubona izinhlelo zokusebenza ezintsha ze-AIOps ekuphathweni kwe-IT. Ngamandla ayo okuguqula indlela esiphatha ngayo izinhlelo ze-IT, i-AIOps iyinkambu ethokozisayo nekhula ngokushesha okufanele ihlale iso.

Have Questions?

Let's Talk

Contact me on LinkedIn