A simple GPT bot designed to be installed on a server with logs that might need to be analysed for errors. The bot's config contains a list of services and a log file for each, which will be used by the bot for analysis.
logparse.mp4
See logparse.yaml.tmpl for an example of the config for the log files.
This bot is built using the langchain framework.
make install
This copies the bot's config as well as the bot script to the
default config directory: ~/botmand-engines.
Edit ~/botmand-engines/logparse.yaml and set the OPENAI_API_KEY variable.
Check OpenAI documentation on signing up
for the API and retrieving your API
key.
Also configure the list of LOG_xxx parameters for the list of services to be
monitored.
If botmand is already running, make it reload its engines:
pkill -HUP botmand
The bot only listens to messages directed at it in specific channels (by
default, #logparse). See ~/botmand-engines/logparse.yaml after installation
to configure this.
To interact with it, ask it a question about one or more of the services configured in the config file. For instance, try "@Botman Does the webserver look okay?"