A practical toolkit of PowerShell, Python, and SQL*Plus automation for Oracle DBA operations — scheduled health checks, RMAN and capacity reporting, alert-log parsing, log collection and rotation, and HTML/email report delivery across many databases. The repeatable tooling that turns manual morning checks into an unattended, alert-driven routine.

This is the Database Reliability Engineering side of the DBA role: less typing the same checks every day, more building automation that runs itself and pages a human only when something needs attention.

A DBA managing several databases can't hand-run the same checks on each one every morning — it doesn't scale, and it's exactly where things get missed. This toolkit packages those checks as scripts that:

• run on a schedule (Windows Task Scheduler or Linux cron / systemd),
• query many targets from one config-driven run,
• classify findings OK / WARN / ALERT against tunable thresholds,
• render a single HTML report and email it (optionally only on alert), and
• exit non-zero on real problems so a scheduler can branch or page.

Everything authenticates through an Oracle wallet and OS environment variables — no passwords in the repo, config, or command line (see the security doc).

All scripts, configs, sample outputs, and docs are fully sanitized. They contain no real hostnames, IPs, ports, service names, DBIDs, SIDs, schema/user names, credentials, or employer/company data. Names such as ORADEMO, SALES_PDB, APPDB01, APPDB02, monitor_user, and demo.example.local are fictional examples. Only the *.example.json config templates are committed — real configs and any wallet/secret are git-ignored. Test in a non-production environment and apply under your own change-control process.

oracle-automation-toolkit/
├── README.md
├── LICENSE
├── .gitignore
├── banner.png
├── powershell/          # 6 PowerShell automation scripts
├── python/              # 5 Python report/parsing scripts (stdlib only)
├── sqlplus/             # 3 spoolable SQL*Plus wrapper queries
├── config/              # 3 *.example.json templates (copy + edit; real ones ignored)
├── docs/                # scheduling, configuration, troubleshooting, security
├── sample_outputs/      # fictional, sanitized example output for the scripts
└── screenshots/         # sanitized captures — see "Operational Screenshots"

Sanitized, annotated sample output for the scripts is in sample_outputs/.

   db_targets.json ──┐
   thresholds.json ──┤    ┌─ sqlplus/*.sql (spoolable, pipe-delimited) ─┐
   email_settings ───┘    │                                            ▼
        │                 │   PowerShell / Python parse + classify (OK/WARN/ALERT)
        ▼                 │                                            │
   scheduler (Task Sched / cron / systemd)                            ▼
        │                                          HTML report  ──►  SMTP email
        └──────────────► runs unattended ──────────────────────►  (alert-only optional)
   credentials: Oracle WALLET (/@alias) + env vars   (never in the repo)

1. Configure (copy the templates; real names are git-ignored):

cp config/db_targets.example.json    config/db_targets.json
cp config/thresholds.example.json    config/thresholds.json
cp config/email_settings.example.json config/email_settings.json

Set up a wallet alias per target so sqlplus -s /@<alias> works without a password (see Security and Credential Handling).

2. Run on demand:

# Windows / PowerShell
.\powershell\Invoke-OracleHealthCheck.ps1 -Target APPDB01 -Html
.\powershell\Send-HealthCheckReport.ps1 -AlertOnly

# Linux / Python
python3 python/oracle_health_report.py --config config --out reports
python3 python/tablespace_threshold_report.py --config config
python3 python/parse_alert_log.py /path/alert_ORADEMO.log --hours 24

Windows Task Scheduler (daily 07:00):

$a = New-ScheduledTaskAction -Execute 'pwsh.exe' `
  -Argument '-NoProfile -File "C:\oracle-automation-toolkit\powershell\Send-HealthCheckReport.ps1" -AlertOnly'
$t = New-ScheduledTaskTrigger -Daily -At 7:00am
Register-ScheduledTask -TaskName 'Oracle Health Report' -Action $a -Trigger $t -User 'DEMO\oracle_svc'

Linux cron (daily 07:00 + 5-minute blocking probe):

0 7 * * *   /opt/oracle-automation-toolkit/run_health.sh >> /var/log/ora_health.log 2>&1
*/5 * * * * /opt/oracle-automation-toolkit/python/blocking_session_report.py --config /opt/oracle-automation-toolkit/config >> /var/log/ora_block.log 2>&1

Full setup in Windows Task Scheduler Setup and Linux Cron Setup.

Scripts describe intent; these show the work. Real terminal runs and the rendered report from a sanitized lab (APPDB01 / APPDB02, ORADEMO), tracing the loop: run across many databases → classify → render → schedule → deliver. Every value is fictional — no hostnames, services, or company data. The point is the engineering judgment: make checks repeatable, threshold-driven, unattended, and credential-safe.

What it demonstrates. One command checks two databases, classifies each finding OK/WARN/ALERT, returns a per-target severity rollup, and writes an HTML report.

Why a senior DBA cares. This is automation, not a one-off script — it's config-driven (add a target in JSON, not a code change) and severity-aware, which is what makes it safe to schedule and trust.

Action that should follow. Investigate the WARN (here, 3 invalid objects — recompile and re-check); schedule the script to run unattended and email on alert.

What it demonstrates. The generated self-contained HTML report — color-coded OK/WARN/ALERT across both databases — produced identically by the PowerShell and Python paths.

Why a senior DBA cares. Raw script output isn't a deliverable; a clean, skimmable report is. A team can read it in five seconds and know exactly what needs attention — and it's one file, easy to email or archive.

Action that should follow. Wire it into the daily email (Send-HealthCheckReport.ps1 or the cron Python job) and trend the WARN/ALERT rows for capacity planning.

What it demonstrates. The health report registered as a daily 07:00 task under a dedicated low-privilege service account — plus a frequent listener probe and a weekly log rotation.

Why a senior DBA cares. "Runs unattended" is the entire point of automation. Running under a dedicated service account (not a personal admin login) with no human in the loop is what separates a script from a reliability system.

Action that should follow. Confirm the task runs whether the user is logged on or not, set a max runtime, and check Task Scheduler / cron history for missed runs. See Windows Task Scheduler Setup.

What it demonstrates. A multi-MB alert log reduced to a short, actionable summary for a time window — counts, top ORA- codes, samples — with a non-zero exit when an archiver/corruption/tablespace event appears.

Why a senior DBA cares. Nobody reads the whole alert log daily; automation that surfaces only what matters — and signals it via exit code — is how you catch an ORA-00257 the morning it happens instead of during the next outage.

Action that should follow. Schedule it; branch on the exit code to page specifically on archiver/corruption/space events. Pair with Collect-OracleLogs.ps1 to bundle evidence when something fires.

What it demonstrates. The report arriving in the team inbox with a severity in the subject line — detection → report → delivery, fully automated.

Why a senior DBA cares. A report nobody sees is worthless. Delivery (with an alert-only option to avoid noise) is what makes the automation operational — and the SMTP password came from an environment variable, never the repo.

Action that should follow. Use -AlertOnly once the baseline is quiet so the team is paged only on WARN/ALERT, and route alerts to the on-call channel.

All screenshots are fully sanitized and fictional. APPDB01, APPDB02, ORADEMO, SALES_PDB, monitor_user, demo.example.local, and all values are illustrative demo data — no production, employer, or confidential information is shown. Each capture mirrors the annotated output in sample_outputs/.

• oracledb (python-oracledb) thin-mode option to query without a local client, alongside the SQL*Plus path.
• Pushgateway / Prometheus exporter so metrics feed an existing monitoring stack, not just email.
• Slack / Teams webhook delivery in addition to SMTP.
• Trend store — append capacity metrics to a small table/CSV for growth charts.
• Self-test / --dry-run mode that validates config and connectivity without querying.
• Data Guard & RAC report sections wired into the same scheduled run.

Part of a broader Oracle DBA portfolio (github.com/uusen01): oracle-rman-scripts · oracle-patching-runbooks · oracle-performance-tuning · oracle-cdb-pdb-administration · oracle-health-checks · oracle-dba-runbooks · oracle-goldengate-configs · oracle-dataguard-runbooks · oracle-rac-administration.

Released under the MIT License — free to use, adapt, and share with attribution.