Viscosity
logo-black
  • Data
    • Database Services
      • Performance Assessment
      • Proactive Healthcheck
      • Database Migration & Consolidation
      • Performance Tuning
    • Microsoft Services
      • Microsoft Azure Services
      • Microsoft SQL Server Services
      • Microsoft Gold Azure Partner
    • Oracle Services
      • Oracle Database 19c
      • RAC
      • GoldenGate
      • Data Guard
      • Oracle & SQL Database
    • Viscosity Remote Services
  • Apps
    • App Dev Services
    • Oracle APEX
    • Viscosity AMP
    • Viscosity-ai
    • Shane-ai
  • Cloud
    • Cloud Services
    • Hybrid Cloud
    • Viscosity Edge Cloud
    • Virtualization & Cloud Expertise
    • Microsoft Azure Services
  • Infrastructure
    • Infrastructure Services
    • Exadata
      • Exadata Resale & Services
    • Oracle Database Appliance
      • Oracle Database Appliance Resale & Services
      • ODA Health Checks
    • Zero Data Loss Recovery Appliance
    • VMware to KVM Migration
  • Events
    • Upcoming Events
    • Virtual Training
    • 2025 Past Events
    • 2024 Past Events
    • 2023 Past Events
    • 2022 Past Events
  • About
    • About Us
    • News
    • Blogs
    • Publications
    • Contact
Hit enter to search or ESC to close
ENGAGE WITH US

Performance Tuning , OraPub , AI , AI bot , performance analysis

Improve Your OraPub Performance Q&A Bot Experience: Tips and Tricks

By Craig Shallahamer
July 05, 2023

The OraPub Oracle Performance And Internals Q&A Bot has been available for several weeks. The questions everyone has asked, and the bot's responses have been amazing and sometimes not all that helpful.

 

Besides improving the underlying AI, what I have learned from analyzing everyone's questions and the bot's responses is that most questions could have resulted in better responses!

 

How could the inputs have resulted in a better bot response?

I noticed two main areas. First, the bot has limitations; there are certain tasks it is simply not designed to perform. Second, I observed that by making slight changes to the inputs provided to the bot, the responses could have been significantly improved.

 

Generally speaking, all the tips below revolve around helping the bot better understand your intentions. In other words, try to make it easy for the bot to comprehend your query, and you will likely receive a more helpful response.

 

So everyone knows, the bot receives two main inputs: your SITUATION and your QUESTION/REQUEST.

 

Bot Limitations 

  • The bot is unable to provide a specific contextual source, such as a URL, the name of a book, or a technical paper. Instead of inquiring about the source's location, ask about the specific content or information within the document.
  • The bot does not retain memory of your previous submissions, including the most recent entry you made.
  • The bot does not generate code or provide scripts.

Top Ten Tips For A Better Bot Experience 

  1. Want a different and perhaps better response? Simply re-submit. 
  2. The bot requires both your situation and your question. The SITUATION refers to the facts or circumstances you are facing, which are often quantifiable. These are not questions or requests. On the other hand, the QUESTION is your specific request or inquiry directed towards the bot, usually ending with a question mark. However, questions can also be formulated as commands, such as "I need a step-by-step solution."
  3. Don't start with the maximum level of detail. Consider starting with the medium level. Then, if the bot understands your question and is able to provide a pretty good response, then increase the detail level.
  4. A slight change in your input can result in a very different and perhaps more useful response. For example, instead of asking, "What does sql execute elapsed time being higher than db cpu time mean?" try something like, "How can elapsed time be larger than 'db cpu' time?" or "How can elapsed time be higher than cpu time?" or "How can elapsed time be higher than foreground process cpu consumption?"
  5. When using an acronym such as ASH, spell out the words in quotes and also provide the acronym.
  6. Trying framing the SITUATION to provide context for your QUESTION. For example:
    • Situation: I need to restore my database using RMAN from a tape backup.
    • Question: What are the steps I must follow to restore a database from a tape backup using RMAN?
  7. Instead of "100mils," try "100M," "100 million," or "one hundred million."
  8. While the bot's response may appear strange, unusual, or even incorrect, it can often be thought-provoking and spark creative and powerful solutions. When you provide an honest and genuine situation along with an honest question, lean into the bot's response. Many times, the bot can make connections and associations that we may have never considered or been aware of through our own education or experiences.
  9. Tell the bot what you need. For example, if you need detailed specific steps, ask for that.
  10. If you want specific information, make sure to request it while also providing specific details about the situation.
    • Situation: 80% of "Database Time" (DB Time) is Oracle foreground process CPU consumption (DB CPU). Operating system CPU utilization is 87%.
    • Question: What specific type of SQL or PL/SQL would be involved in this situation?

If you have any questions or need some help, feel free to email me.

 

Enjoy the bot! It's just the beginning.

 

All the best in your Oracle tuning and machine learning work,

 

Craig

All posts
About Author
Craig Shallahamer

Craig is a co-founder of OraPub and Viscosity's Applied AI Scientist. Having worked with Oracle technology since 1989, Craig Shallahamer is a leader in the fields of machine learning, artificial intelligence and Oracle database performance tuning. Craig has extensive experience in constructing and teaching predictive modeling methods, notably developing a Reinforcement Machine Learning bot in 1990. He has launched several specialized generative AI conversational assistants, each with distinct personalities and capabilities. As an Applied AI Scientist at Viscosity and the founder of OraPub, Craig is also the author of two acclaimed books: "Oracle Performance Firefighting" and "Forecasting Oracle Performance". He has received recognition as an Oracle ACE Director for his valuable contributions to the Oracle community through his technical expertise and leadership. Craig is an active participant in Oracle user groups, frequently presenting at conferences and serving as a board member and volunteer.

You might also like
Introducing the New AI-Powered Oracle Performance Q&A Bot: A Game-Changer by Craig Shallahamer
Introducing the New AI-Powered Oracle Performance Q&A Bot: A Game-Changer by Craig Shallahamer
July 05, 2023
How to Tell if the IO Subsystem Reads Are Struggling
How to Tell if the IO Subsystem Reads Are Struggling
July 05, 2023
SUBMIT YOUR COMMENT
logo for footer

Viscosity's core expertise includes:

Data Transformation, Emerging Technology, High Availability & Scalability Solutions, Cloud Migrations, Performance Tuning, Data Integrations, Machine Learning, APEX Development, and Custom Application Development.


Solutions

Resources

Partnerships

Careers

Clients

 

Contact
Email: sales@viscosityna.com

Telephone:
(469) 444-1380

Address:
3016 Communications Pkwy Suite 200, Plano, TX 75093

Copyright 2025. All Rights Reserved by Viscosity North America.