A tool to send you alerts from your AWS instances
Install
pip install git+https://github.com/MantisAI/bell.git@main
Use bell ๐ to get alerts that your AWS instance is on
bell --webhook-url SLACK_WEBHOOK_URL
Use bell ๐ to get alerts about when a command starts, fails or finishes
export SLACK_WEBHOOK_URL=xxx
bell python -c "print('hello')"
Note that you can pass the slack webhook URL either as a command line parameter
--webhook-url
or as an environment variable SLACK_WEBHOOK_URL
You can capture the output of the process by passing the --capture-output
option. This will then get printed to slack.
bell --cpature-output python -c "print('hello')"
The results will be printed in slack as:
๐๏ธ command python -c print('hello') started ๐
hello
๐๏ธ command python -c print('hello') finished ๐ฅ
You can read how to setup a slack webbook here
This tool works well with a scheduler like cron. Here is an example cron command which sends an alert every 2 hours
SLACK_WEBHOOK_URL=xxx
0 */2 * * * bell
Note that cron's environment is different to your user or root environment. See troubleshooting if you have problems setting up bell to work with cron.
usage: bell [-h] [--webhook-url WEBHOOK_URL]
optional arguments:
-h, --help show this help message and exit
--webhook-url WEBHOOK_URL
incoming slack webhook url
By default bell sets the logging level at INFO. For debugging purposes, it is more helpful
to set it at DEBUG. You can do that via the environment variable LOGGING_LEVEL
by running
export LOGGING_LEVEL=DEBUG
In order for cron to work it needs to have access to both bell, python packages and AWS credentials. For this reason we recommend you install a cron job under your user account and not root, otherwise you might need to install the tool with sudo and ensure root has access to AWS credentials.
Even in that case as it turns out cron has a different more minimal PATH
than what you see in the terminal if you run echo $PATH
. For that reason, you should
probably define the same PATH in your cron job as the one you see when you test the
tool.
We assume you have AWS credentials setup either through environment variables, aws configure or proper IAM roles. We use boto under the hood to query AWS, see its documentation around credentials here