79 lines (59 loc) · 2.67 KB

Ringmaster Handlers

Each file in your stack directory is processed by a different handler. Handlers are just Python functions that are executed when a filename matches the handler pattern.

Ringmaster supports these handlers:

*.sh

Normal bash scripts
Each variable in databag exposed as environment variables
Put values in databag by writing JSON to $intermediate_databag_file
script will be run with $1 == $up_verb if stack is creating or $1 == $down_verb if stack is being destroyed

*.cloudformation.yaml

Normal cloudformation in yaml format
Parameters are converted to snake_case and looked up from databag
Outputs are converted to snake_case and added to databag

*.remote_cloudformation.yaml

Remotely hosted cloudformation scripts to get around 51200 byte upload hard limit
Handled the same as local cloudformation scripts
A copy of the remote file will be downloaded for your own records and for parameter pre-processing

*.kubectl.yaml

Databag variables are available and can be inserted as ${variable_name}
Ringmaster pre-process the file to substitute variables and saves the output as *.kubectl.processed.yaml
*.kubectl.processed.yaml files are just normal kubectl files

kustomization.yaml

Processed with kubectl (apply|delete) -k
Not changed by ringmaster in any way
Be sure to also download any required files

*.ringmaster.py

Rudimentary plugin system
Normal python scripts
Top level variable databag will be set with the current databag
To make new data available to other stages, just add it to databag
Ringmaster runs your script like this:
1. set databag
2. call main() with argument verb which indicates whether we are creating or destroying

*.snowflake.sql

Bunch of SQL commands to run against snowflake
Configure snowflake credentials at ~/.ringmaster/snowflake.yaml
Placeholders will be substituted and the result saved to file.snowflake.processed.sql
Each line must end with ;
Lines starting -- will be discarded
One statement per line, but long statements can be split over multiple lines

*.snowflake_query.sql

ONE sql command to run which must be a SELECT statement
Configure snowflake credentials at ~/.ringmaster/snowflake.yaml
Placeholders will be substituted and the result saved to file.snowflake_query.processed.sql
Entire processed file will be executed as-is
Columns in the results will be added to databag using the column name. Use SQL AS to set databag name, eg: SELECT x AS the_name_for_databag

helm_deploy.yaml

Install helm repo and deploy directly
Repeatable deployments artifactory integration or similar