search

epam / badgerdoc

Apache License 2.0
33 stars 32 forks source link

[CRITICAL] Annotation cannot be done for a long document #615

Open MarinaVorobeva21 opened 1 year ago

MarinaVorobeva21 commented 1 year ago

Precondition: There is a document with more than 300 pages.

STR: Go to Upload Wizard in Documents Upload a document with more than 300 pages Create Extraction and annotation (Extensive coverage Validation type) Start the job Try to annotate the document

AR: Text annotations and others cannot be done. https://github.com/epam/badgerdoc/assets/98399603/4d48de09-b95c-417a-9d48-d93d81cb5ccf

ER: Annotation is created successfully.

serereg commented 1 year ago

Calling separate endpoint within cluster .../api/v1/convert/docs#/pdf/import_pdf_pdf_import_post now gives the same 504 upstream request timeout but all pages (>300 pages) json on minio processes well.

It turns out that the problem is in this "intermediate" error. The error comes after 60 s even though most timeouts were made more than 1000 s. So it seems, the problem is in Load Balancer timeout.

khyurri commented 1 year ago

When user annotates huge document, UI became very slow. We need to check reason and fix it