v0.15.0

New: Maintenance

• Equipment as a tree (line › machine › component) with its criticality, the product zone, a documents folder and a QR to stick on the machine that opens the breakdown report with it chosen.
• Anyone but guests reports a breakdown, readers and the plant tablets' username + PIN accounts too (a big card on their home). Whoever only reports sees their own reports.
• Work orders with their cycle (reported, planned, in progress, on hold, done, closed), each technician's time, spare parts, lockout/tagout warning and the hygiene release quality gives before closing an order on product-zone equipment (IFS 4.17.4).
• Report and close by talking: with Bian on, a Record button; the audio is transcribed on your own server with Whisper (optional audio compose profile, WHISPER_URL) or by the browser's dictation, and the AI fills in the report or the closing (cause, solution, time, parts) for the person to check. Audio is kept 30 days.
• Suggest a resolution: the AI reads the symptom, the resolved breakdowns of that equipment and of the same model, and the manuals the person can see, and proposes causes with steps, citing its sources.
• Preventive and calibration plans by time or by meter usage, with tasks to tick; their work orders are created by themselves. Calibration plans show on Quality › Prerequisites.
• Quality link: a monitoring record out of limits that opens a nonconformity also opens an urgent work order on the equipment it names; the nonconformity follows its work order.
• Spare parts from the warehouse system (SGA): integration tokens (owner only, shown once, revocable) and an API to send the catalogue and stock (PUT /api/v1/integrations/maintenance/parts) and read the parts used (GET /api/v1/integrations/maintenance/consumption).
• Summary: availability, MTBF, MTTR, downtime, preventive versus corrective, preventive compliance, open work orders and the equipment with the most breakdowns, for the plant or a line, with CSV export.
• Turned on per company in Administration › Security › General, with three new capabilities: manage, work on and view maintenance.

Many people at once

• The request limit now counts per person when signed in (RATE_LIMIT_PER_USER, 300 a minute), so a whole plant behind one IP no longer gets "too many requests"; without a session it's per IP, and everything signed in from one IP has a cap (RATE_LIMIT_PER_SHARED_IP, 6000). Sign-ins per IP go to 500 every 15 minutes (LOGIN_LIMIT_PER_IP); the 5 tries per account and the lockout stay.
• Postgres connections per process are configurable (DATABASE_POOL_MAX, 20 by default, was 10), and API_PROCESSES runs the API in several processes, one per core. A saturated database now answers "busy, try again" instead of an error.

Upgrading
Run update.ps1 (Windows) or update.sh (Linux) from your spiderfolio folder. This version changes the database: the app applies the changes when it starts, and the update script backs up first. To report and close by talking with audio on your server, add audio to COMPOSE_PROFILES and set WHISPER_URL=http://whisper:9000.