administration — admin only
System
Pipeline state, scheduled tasks, user management and Upstox API integration.
every pipeline, checked each minute
Supervisor
Each process is judged by what it produced, not by whether its task reported success — a job that runs on schedule and writes nothing is the failure this is built to catch.
recent repairs
diagnostics — data & services
System Health Check
Reports what is in the database, not whether a job exited cleanly — a task that finds nothing to do also reports success.
eod (1-day) — nse, options excluded
Coverage by Asset Class
Run the check.
bars — by interval
Ingested Bars
—
1-minute — session completeness
Intraday Sessions
A full NSE cash session is 375 one-minute bars.
—
backfill — per instrument
Backfill Progress
done — history stored · empty — Upstox returned no candles, normal for far-dated or illiquid contracts · failed — the request errored
—
downstream — derived datasets
Dataset Freshness
—
accounts — all users
User Management
Login tokens are shown so they can be issued to users. Anyone with a token, an email and a password can sign in — treat this list as credential material.
Loading…
authentication — upstox
Authentication Status
Loading…
api integration — credentials
Upstox API Credentials
Stored server-side and used ahead of environment values. Leave a field blank to keep its current value — a blank submit never wipes working credentials. Secrets are never sent back to this page.
pipeline — data freshness
Process Status
Loading…
two schedulers
Scheduled Tasks
host — cron
Host Jobs
Backups, log and container collectors, and the on-demand backup queue. These run on the host because the worker is containerised and can reach neither pg_dump, host journald, nor the Docker socket — and granting it any of those to unify the two schedulers would put root-equivalent access inside the web application.
Loading…
application — celery beat
Pipeline Tasks
Read live from the running scheduler, so this cannot drift from what beat is actually executing.
Loading…
audit trail — portal and host
Activity Log
User activity is what happened through the portal. System activity is the layer beneath it — SSH, sudo and container lifecycle on the host, where most real change actually happens.
portal — role changes, tokens, password resets
Loading…
host — ssh, sudo, container lifecycle
The layer beneath the portal. Collected on the host every 5 minutes — most real change happens over SSH, which the web UI never sees.
Loading…
retention — 3 days live, 3 years archived
Activity Storage
disaster recovery
Backup & Restore
what a backup contains
Coverage
history — newest first
Backup Runs
Loading…
procedure
Restoring
host — collected every minute
Docker & Containers
this application
Stack Containers
Loading…
sharing this host — not managed by this project
Other Containers
A live n8n/Traefik stack runs on this machine and predates this project. It is shown because it competes for the same CPU, memory and disk — not because it is ours to touch.
storage
Disk & Images
containers & host journal — warnings and errors
Syslogs
Only WARN and above is stored. Gunicorn logs every request at INFO; keeping that would add millions of rows a week duplicating what nginx already records, and would bury the lines that matter.
Loading…