I-VictoriaMetrics, uchungechunge lwesikhathi lwe-DBMS oluhambisana ne-Prometheus, luvulekile

Vula ikhodi yomthombo I-VictoriaMetrics - i-DBMS esheshayo futhi eyingozi yokugcina nokucubungula idatha ngendlela yochungechunge lwesikhathi (isikhathi sakha irekhodi kanye nesethi yamanani ahambisana nalesi sikhathi, isibonelo, atholakala ngokuvota ngezikhathi ezithile kwesimo sezinzwa noma ukuqoqwa kwezinzwa amamethrikhi). Iphrojekthi incintisana nezixazululo ezinjengokuthi I-InfluxDB, I-TimescaleDB, Thanos, cortex и Uber M3. Ikhodi ibhalwe ngolimi lwe-Go futhi isatshalaliswa ngu ilayisensi ngaphansi kwe-Apache 2.0.

Izinzuzo nezici ze-VictoriaMetrics:

• Kulula ukuyisebenzisa. Kuyifayela elilodwa elisebenzisekayo elinezilungiselelo ezincane ezidlule emugqeni womyalo lapho uqala. Yonke idatha igcinwa kuhla lwemibhalo olulodwa, olucaciswe ekuqaleni kusetshenziswa ifulegi elithi “-storageDataPath”;
• Umbuzo wosekelo lolimi I-PromQL, esetshenziswa ohlelweni lokuqapha Prometheus. Imibuzo engezansi ye-PromQL futhi eminye iyasekelwa amakhono andwe, njengenkulumo ethi "offset", amaphethini angaphakathi kwezitatimende "WIDTH", "uma" kanye "nokuzenzakalelayo", imisebenzi eyengeziwe, kanye nekhono lokufaka amazwana;
• Ingasetshenziswa njenge ukugcinwa kwedatha yesikhathi esideexhunywe ku-Prometheus kanye UGrafana.
• Ukutholakala kwemodi yokugcwalisa emuva ukuze kulayishwe idatha yomlando;
• Isekela amaphrothokholi ahlukahlukene okudlulisa idatha, okuhlanganisa I-Prometheus API, Ithonya, I-Graphite и I-OpenTSDB. I-VictoriaMetrics ingasetshenziswa njengokumiselela okusobala kwe-InfluxDB futhi ingasebenza nabaqoqi abahambisana ne-InfluxDB njenge-Telegraf;
• Ukusebenza okuphezulu nokusetshenziswa kwezinsiza eziphansi kuqhathaniswa ngezinhlelo ezincintisanayo. Kwezinye izivivinyo, i-VictoriaMetrics idlula i-InfluxDB ne-TimescaleDB izikhathi ezingafika kwezi-20 lapho yenza imisebenzi yokufaka nokulanda. Lapho wenza imibuzo yokuhlaziya, inzuzo eqhathaniswa ne-DBMS PostgreSQL ehlobene ne-MySQL ingaba izikhathi ezisukela kweziyi-10 kuye kweziyi-1000.
  
  
  

• Kuyatholakala ithuba icubungula inombolo enkulu kakhulu yochungechunge lwesikhathi oluhlukile. Lapho ucubungula izigidi zochungechunge lwesikhathi oluhlukile, zisebenzisa i-RAM engaphansi kwezikhathi eziyi-10 kune-InfluxDB.
• Izinga eliphakeme lokuminyanisa idatha ekugcinweni kwediski. Uma kuqhathaniswa ne-TimescaleDB, ingakwazi ukufaka amarekhodi afika izikhathi ezingu-70 ngenani elifanayo lokulondoloza;
• Ukutholakala kokulungiselelwa kwesitoreji nge-latency ephezulu kanye nenani eliphansi lemisebenzi yokufaka/yokukhiphayo ngomzuzwana (ngokwesibonelo, ama-hard drive kanye nesitoreji samafu i-AWS, i-Google Cloud ne-Microsoft Azure);
• Isistimu yokusekelayo elula esekelwe izifinyezo;
• Ukutholakala kwezindlela zokuvikela ubuqotho besitoreji emonakalweni wedatha, isibonelo, uma kwenzeka ugesi unqamuka (indawo yokugcina inefomu isihlahla esakhiwe ngogodo ngokuhlanganisa);
• Ukusetshenziswa ngolimi lwe-Go, okunikeza ukuhwebelana phakathi kokusebenza nokuba yinkimbinkimbi kwekhodi uma kuqhathaniswa neRust kanye ne-C++.
• Amakhodi omthombo anikeziwe izinguqulo zeqoqo, esekela ukukala okuvundlile kumaseva amaningi futhi ibonise phezulu okuphansi. Izici zokutholakala okuphezulu ziyatholakala.

Source: opennet.ru