Dlp Oncall Tracker

Common Queries

# Raw messages for a date
sqlite3 /root/code/tools/dlp_oncall_tracker/dlp_oncall.db \
  "SELECT ts, user_name, text FROM messages
   WHERE date(fetched_at) = '2026-03-24'
   ORDER BY ts;"

# Messages for last 24 hours
sqlite3 /root/code/tools/dlp_oncall_tracker/dlp_oncall.db \
  "SELECT ts, user_name, text FROM messages
   WHERE fetched_at >= datetime('now', '-1 day')
   ORDER BY ts DESC;"

# Messages for a date range
sqlite3 /root/code/tools/dlp_oncall_tracker/dlp_oncall.db \
  "SELECT date(fetched_at) as day, COUNT(*) as msgs
   FROM messages
   WHERE fetched_at >= datetime('now', '-7 days')
   GROUP BY day ORDER BY day DESC;"

# Run health check
sqlite3 /root/code/tools/dlp_oncall_tracker/dlp_oncall.db \
  "SELECT run_at, status, message_count, error FROM run_log ORDER BY run_at DESC LIMIT 10;"

# Jira tickets seen recently
sqlite3 /root/code/tools/dlp_oncall_tracker/dlp_oncall.db \
  "SELECT ticket_id, title, status, assignee, priority FROM jira_tickets ORDER BY fetched_at DESC;"

Agent Workflow for "Today's Issues"

1. Query messages for today:

   SELECT ts, user_name, text FROM messages
   WHERE date(fetched_at) = date('now')
   ORDER BY ts;
   

2. If no rows → run python3 /root/code/tools/dlp_oncall_tracker/main.py via Bash, then re-query
3. Read the messages and summarize inline — categorize them yourself using the DLP Oncall Categories below. Do NOT invoke a subprocess or external tool for summarization.

DLP Oncall Categories

When summarizing messages, classify issues into these buckets:

For each category found, report:

• Severity: critical / high / medium / low
• Incident count: number of distinct incidents
• Plane: DP, MP, both, or unknown
• Open: whether the issue appears resolved in the thread
• Summary: 1-2 sentence description
• Jira tickets: any [A-Z]+-\d+ patterns mentioned

Manual Run (Fetch Missing Days)

python3 /root/code/tools/dlp_oncall_tracker/main.py

Idempotent — only fetches days not yet in run_log. Safe to re-run.