Monitor — django_monitor
django_monitor captures server-side errors, slow queries, background-task
failures and browser events, writes each as a JSON line through the standard
Django-CFG logging pipeline, and (when Telegram is configured) batches error
events into Telegram alerts.
Zero-config. There is no MonitorConfig — nothing to switch on. Capture
hooks connect automatically at startup. Telegram alerts turn on by themselves
the moment a TelegramConfig is present on your DjangoConfig; with no Telegram
configured, capture still records events, it just doesn’t send alerts.
How It Fits Together
Every event lands in logs/djangocfg/monitor.log (daily rotation, 30-day
retention — the same file infrastructure as all other Django-CFG logs).
Error-level events additionally join a Telegram batch when Telegram is set up.
What Gets Captured
| Source | Event type | Alerts? |
|---|---|---|
got_request_exception signal | UNHANDLED_EXCEPTION | yes |
logging.Handler (ERROR+) | LOG_ERROR | yes |
execute_wrapper (slow DB query) | SLOW_QUERY | over threshold |
RQ exception_handler | RQ_FAILURE | yes |
capture_exception() | SERVER_ERROR | yes |
capture_message() | LOG_ERROR | yes (by level) |
Browser events (/cfg/monitor/ingest/) | FRONTEND_ERROR, … | error-level only |
Alerts are batched: the first occurrence of a new fingerprint is sent immediately, and repeats within the window are folded into a single 60-second digest so a crash loop can’t flood your chat.
Quick Start
There is nothing to configure for capture. To receive alerts, add a
TelegramConfig:
# djangoconfig.py
from django_cfg import DjangoConfig, TelegramConfig
class MyConfig(DjangoConfig):
telegram: TelegramConfig = TelegramConfig(
bot_token="${TELEGRAM_BOT_TOKEN}",
chat_id="${TELEGRAM_CHAT_ID}",
)That’s the whole setup — errors now flow to monitor.log and to Telegram.
Manual capture (optional)
from django_cfg.modules.django_monitor import capture_exception, capture_message
try:
process_payment(order)
except Exception as e:
capture_exception(e, url="/api/orders/", http_method="POST")
capture_message("payment gateway slow", level="warning", extra={"latency_ms": 3200})Both calls are fire-and-forget — they never raise.
Check status
python manage.py monitor_statusReports whether alerts are on (i.e. whether Telegram is configured) and
summarizes today’s events from monitor.log.
Browser events
The frontend counterpart is the @djangocfg/devtools
package. It captures JS errors, console errors and failed requests in the
browser and posts them in batches to POST /cfg/monitor/ingest/, where they are
enriched and written to monitor.log like any server event (error-level ones
also alert). It ships an in-app debug panel for inspecting the same feed live.
What’s Next
@djangocfg/devtools browser package and debug panelManagement Commandsmonitor_statusTAGS: django_monitor, error-tracking, telegram, logging, zero-config
Django-CFG