Wụga na isi ọdịnaya
&Sageio

MBUGHARỊ ỤLỌ ỌRỤ

Sageio n'ime akụrụngwa nke gị.

Maka ụlọ ọrụ na-enweghị ike itinye data nzukọ n'igwe ojii ọtụtụ onye. Sageio na-arụ ọrụ na VPC gị, ụlọ data gị, ma ọ bụ igwe ojii nzuzo gị — n'okpuru oke nchekwa gị, na ndekọ nyocha gị.

Gwa ndị ire ahịa Budata ntuziaka mbugharị

IHE MERE NJIDE ONWE

Mgbe n'ụlọ bụ naanị azịza.

  • Ebe data bi

    Ndị na-achị, nkwekọrịta onye ahịa, ma ọ bụ amụma n'ime na-achọ ka data nzukọ nọrọ n'ime otu mba ma ọ bụ netwọk. Njide onwe na-etinye data kpọmkwem ebe ọ ga-anọ — ọ dịghị ebe ọ na-ekwesịghị ịnọ.

  • Ọrụ nrube isi

    Ụlọ ọrụ n'okpuru FINRA, HIPAA, ITAR, ma ọ bụ usoro mpaghara dabara na-eche mkpa nyocha nke SaaS adịghịkarị emejupụta. Ịrụ Sageio na gburugburu gị na-eme ka ị nọgide na-achịkwa ihe akaebe — na ajụjụ ndị nyocha.

  • Oke nchekwa dị adị

    Ụfọdụ ụlọ ọrụ etinyela afọ na nkewa netwọk, CA n'ime, na ngwaọrụ zero-trust. Njide onwe na-eme ka Sageio nweta oke ahụ kama ịgbawa oghere site na ya.

OWUWU

Ihe na-arụ ọrụ ebe.

Sageio njide onwe bụ ngwa containerized nke na-arụ ọrụ site na njedebe ruo njedebe n'ime gburugburu gị. Nkọwa AI, nkọwa okwu, nchekwa — niile n'ime oke netwọk gị.

Mbugharị isi gụnyere sava ngwa Sageio, nchekwa data PostgreSQL, nchekwa ihe maka ọdịyo na nkọwa, na ọkwa nkọwa maka nkọwa okwu na ntụgharị. Ndị ahịa nwere ike iweta STT na LLM nke ha (Deepgram, Azure OpenAI, Google Vertex, ma ọ bụ ihe yiri ya) ma ọ bụ rụọ nkpokọ nkọwa ntụaka Sageio na GPU mpaghara. Ohere netwọk apụ apụ chọrọ naanị maka nkwenye ikike na telemetry nhọrọ, nke abụọ nwere ike ịgafe site na proxy onye ahịa na-achịkwa.

Ikpo okwu nzukọ

Meet · Zoom · Teams

Gburugburu gị · oke netwọk

Sava ngwa Sageio

PostgreSQL

Nchekwa ihe

ọdịyo + nkọwa

Nkọwa

STT + ntụgharị

Weta njedebe nke gị (Deepgram, Azure OpenAI, Google Vertex) ma ọ bụ rụọ na GPU mpaghara

Nkwenye ikike + telemetry nhọrọ, site na proxy gị

Njedebe ikike Sageio

Owuwu ntụaka maka mbugharị n'ụlọ. A na-emegharị ya dabere na gburugburu onye ahịa n'oge nkọwa.

MKPA

Ihe gburugburu gị kwesịrị inwe.

Nhazi
Kubernetes 1.27 ma ọ bụ nke ọzọ. Helm 3.12+. A na-akwado Docker Compose maka ihe akaebe na mbugharị otu sava.
Mgbakọ (ọkwa ngwa)
8 vCPU, 16 GB RAM kacha nta kwa node. A na-akwado node abụọ maka ịdị adị elu.
Mgbakọ (ọkwa nkọwa, nhọrọ)
NVIDIA GPU nwere 24 GB VRAM (A10G, L4, ma ọ bụ ihe yiri ya) maka STT njide onwe. Achọghị ya ma ọ bụrụ na ị na-eji njedebe STT / LLM e jikwara.
Nchekwa data
PostgreSQL 14 ma ọ bụ nke ọzọ. Onye ahịa na-ejikwa ma ọ bụ Sageio na-ebuga. Nchekwa 100 GB kacha nta, e tinyere amụma ndebe.
Nchekwa ihe
S3-compatible (AWS S3, MinIO, Ceph, Azure Blob, GCS). E ji ya maka mbupụ nkọwa na oge ndebe ọdịyo.
Mpụta netwọk
HTTPS apụ apụ gaa na njedebe ikike Sageio chọrọ. Okporo ụzọ ndị ọzọ niile nwere ike ịhazi, gụnyere site na proxy onye ahịa.
Njirimara
Onye na-enye njirimara SAML 2.0 ma ọ bụ OIDC. Mmekọrịta a nwalere: Okta, Microsoft Entra ID, Google Workspace, Auth0, Keycloak.
TLS
Asambodo onye ahịa nyere, ma ọ bụ mmekọrịta na CA n'ime. A na-akwado Let's Encrypt maka mbugharị chere ịntanetị ihu.

NRỤNYE

Site na mmalite ruo nzukọ mbụ.

Mbugharị njide onwe nkịtị na-aga izu anọ ruo isii site na ịbịanye aka na nkwekọrịta ruo nzukọ mbụ a sụgharịrị na mmepụta. Sageio na-ekenye injinia ngwọta raara onwe ya maka oge ahụ.

  1. Nkọwa & nyocha gburugburu

    Nnọkọ ọrụ ọnụ na-ekpuchi akụrụngwa gị, onye na-enye njirimara, owuwu netwọk, mkpa nrube isi, na atụmatụ mbugharị. Mmepụta: akwụkwọ atụmatụ mbugharị, ndị otu abụọ bịanyere aka tupu ọrụ nrụnye ọ bụla amalite.

  2. Nrụnye staging

    E tinyere Sageio na gburugburu staging ma ọ bụ pre-production gị. A haziri uru Helm maka nchekwa gị, onye na-enye njirimara, na njedebe nkọwa. A kwadoro nzukọ nnwale site na njedebe ruo njedebe.

  3. Nyocha nchekwa & nrube isi

    Ndị otu nchekwa gị na-arụ nnwale mbanye, nyocha SBOM, na usoro njikwa mgbanwe n'ime ọ bụla. Sageio na-enye akwụkwọ na nzaghachi injinia nye ihe a chọpụtara.

  4. Mbugharị mmepụta & ngafe

    Nrụnye mmepụta, e kwadoro njikọ njirimara, e jikọrọ nlekota na nkpokọ nlebanya gị. Nzukọ mmepụta mbụ na-aga n'akụkụ ngwaọrụ gị dị adị ruo mgbe ndị otu gị nwere obi ike.

  5. Nnyefe & nkwado

    Ọzụzụ onye nchịkwa, ngagharị runbook, na nnyefe akwụkwọ. Nkwado na-aga n'ihu na-agafe gaa n'aka injinia ịga nke ọma onye ahịa e kenyere gị.

NKWADO

Anyị nọnyeere gị site na mmepụta.

Mbugharị njide onwe gụnyere injinia ngwọta tupu, n'oge, na mgbe mbugharị gasịrị. Ọ bụghị ọwa Slack na wiki — mmadụ maara gburugburu gị.

A na-ekenye onye ahịa njide onwe ọ bụla injinia ngwọta a kpọrọ aha maka mbugharị, na injinia ịga nke ọma onye ahịa maka oge nkwekọrịta. Nkwado ọkọlọtọ na-ekpuchi nzaghachi awa azụmaahịa n'ụbọchị ọrụ, na SLA nnabata awa anọ maka okwu mmepụta-daa. Awa gbatịrị na nkwado 24×7 dị dị ka mgbakwunye. Oku nyocha kwa ọnwa atọ na-ekpuchi mmelite ikpo okwu, ndụmọdụ nchekwa, na mmemme nkwalite. A na-ahazi nkwalite mbipụta buru ibu ọnụ ma rụọ ya megide gburugburu staging gị mbụ.

Ọnụ ahịa njide onwe →

Ịchọrọ nke a na usoro gị?

Gwa anyị gbasara gburugburu gị, anyị ga-eziga ajụjụ ọnụ nkọwa. Ọtụtụ nyocha na-emecha n'izu abụọ, gụnyere ihe akaebe na-arụ ọrụ na gburugburu staging gị.

Gwa ndị ire ahịa Budata ntuziaka mbugharị