- Home
- Skills
- Acedergren
- Oci Agent Skills
- Oracle Dba
oracle-dba_skill
- Shell
4
GitHub Stars
2
Bundled Files
2 months ago
Catalog Refreshed
4 months ago
First Indexed
Readme & install
Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.
Installation
Preview and clipboard use veilstrat where the catalogue uses aiagentskills.
npx veilstrat add skill acedergren/oci-agent-skills --skill oracle-dba- metadata.json525 B
- SKILL.md15.8 KB
Overview
This skill provides practical, ADB-focused Oracle Database administration guidance for Autonomous Database on OCI. It targets performance troubleshooting, cost optimization, HA/DR and operational gotchas across ADB versions (19c/21c/23ai/26ai). Use it to follow safe operational patterns, avoid common cost traps, and run SQL_ID-driven debugging workflows.
How this skill works
It codifies decision trees and step-by-step workflows: find slow SQL by SQL_ID, capture and analyze execution plans, run DBMS_SQLTUNE tasks, and interpret v$ views and wait events. It also summarizes exact cost formulas for ECPU/storage, auto-scaling behavior, backup retention impacts, and version-specific features. When OCI API/CLI actions are required, it points to exact CLI command references to avoid guessing syntax.
When to use it
- Troubleshooting one slow query or system-wide performance degradation
- Planning or validating ECPU scaling and auto-scaling limits
- Estimating and optimizing ADB operational costs (ECPU, storage, backups)
- Implementing HA/DR, clones, or version upgrades in ADB
- Auditing and correcting anti-patterns (ADMIN usage, SELECT *)
Best practices
- Never use ADMIN credentials in application code; create least-privilege app users
- Always identify SQL_ID for slow queries and tune specific statements, not global parameters
- Check wait events before scaling ECPUs; scale only after confirming CPU-bound waits
- Use landing zone Terraform modules for secure compartmenting and private endpoints
- Prefer automatic backups for routine retention; set retention on manual backups to avoid ongoing charges
Example use cases
- Find top slow SQL in last hour, capture plan and run a DBMS_SQLTUNE task
- Decide whether to enable auto-scaling by modeling peak ECPU costs and setting a max ECPU cap
- Stop long-idle ADBs: export data and delete instance for >60-day idle to avoid storage/backup charges
- Resolve connection errors by validating wallet, service name and NSG rules
- Upgrade ADB: test in a cloned instance to validate plan changes and new 23ai/26ai features
FAQ
Check v$system_event. If CPU time wait is top and sustained after SQL tuning, scale ECPU. If top waits are I/O or full table scans, optimize SQL or storage IOPS first.
Does stopping ADB remove all costs?
No. Compute stops but storage and backup retention still charge ($0.025/GB/month). For long-term idling, export and delete the database, then restore when needed.