Azure AI for an API Platform
We are managing an API Platform in our company.
We collect metrics in ELK (deployed on premise) and We have a Confluence WIKI page with our public documentation.
We would like to have a Chatbot that can be trained by structured (JSON, CSV) Metrics (Prometheus) stored in ELK and by trained by the unstructured content in our WIKI page.
The Chatbot should be able to answer questions like:
(Source: WIKI page)
– How can I publish an API with mTLS enabled
– How can I authorize a client certificate for my API
– How can I specify the RPS for my API
– and so on and so forth
(Source: ELK)
– Plot the request of the API “XX” for the last month
– Can you predict the API calls trend of the next month for the API “XXX”
– Give me a list of the client IP that accessed the API “YYY” yesterday
– and so on and so forth
In the future We may want the chat bot to be able to do some basic automatic actions like (contacting our self-service API):
– Allow this client certificate to access this API
– Change the RPS for the API ZZZ to 10 rps/s
– and so on and so forth
– What Azure AI Services would you recommend to start looking into? Which AI models? Which Azure resource?
– How can we train and feed the model from ELK? do we have to export the data from ELK daily and store it in an Azure Storage account or we can instruct the specific Azure AI Service to connect to our ELK endpoint (or a proxy API) to fetch the data?
Thank you
We are managing an API Platform in our company.We collect metrics in ELK (deployed on premise) and We have a Confluence WIKI page with our public documentation.We would like to have a Chatbot that can be trained by structured (JSON, CSV) Metrics (Prometheus) stored in ELK and by trained by the unstructured content in our WIKI page.The Chatbot should be able to answer questions like: (Source: WIKI page)- How can I publish an API with mTLS enabled- How can I authorize a client certificate for my API- How can I specify the RPS for my API- and so on and so forth (Source: ELK)- Plot the request of the API “XX” for the last month- Can you predict the API calls trend of the next month for the API “XXX”- Give me a list of the client IP that accessed the API “YYY” yesterday- and so on and so forth In the future We may want the chat bot to be able to do some basic automatic actions like (contacting our self-service API):- Allow this client certificate to access this API- Change the RPS for the API ZZZ to 10 rps/s- and so on and so forth – What Azure AI Services would you recommend to start looking into? Which AI models? Which Azure resource?- How can we train and feed the model from ELK? do we have to export the data from ELK daily and store it in an Azure Storage account or we can instruct the specific Azure AI Service to connect to our ELK endpoint (or a proxy API) to fetch the data? Thank you Read More