Utilities for generating, transforming and transmitting unit cost analytics (UCA) data to the CloudZero API. Visit our UCA documentation to learn more about CloudZero and our enhanced unit cost analytics capabilities.
- Transmit UCA data to CloudZero (using the CloudZero UCA API)
- Generate UCA data
- Convert raw data in CSV form to UCA JSON format
- Tested on MacOS, should probably run on Linux in general
- Python 3.9 or newer
pipxor your favorite method of installing packages from PyPi. Have you considered pipx?
$ pipx install cloudzero-uca-tools
CloudZero UCA tools exist to produce UCA events that can then be transmitted to the CloudZero API for analysis and processing. To use the CloudZero API, you should first obtain an API key from https://app.cloudzero.com/organization/api-keys
$ uca
Usage: uca [OPTIONS] COMMAND [ARGS]...
CloudZero Unit Cost Analytics Toolkit
Options:
--version Show the version and exit.
-c, --configuration TEXT UCA configuration file (JSON)
-dry, --dry-run Perform a dry run, read and transform the data but
do not send it to the API. Sample events will be output to the screen
-k, --api-key TEXT API Key to use
--help Show this message and exit.
Commands:
convert
transmit
generate
A configuration file is typically required to define how UCA telemetry is to be created, converted or transmitted.
The following is a minimal configuration file for transmitting data to the CloudZero API
(if you don't want to specify the API on the command line, otherwise no configuration file
is required for transmit). If you wish to use generate or convert however a configuration file
is required.
{
"version": "1",
"settings": {
"api_key": "<YOUR API KEY HERE>"
}
}
# API Key can also be provided at runtime via the CLI. Get an API key at https://app.cloudzero.com/organization/api-keys
Each command has its own set of command line options that will need to be provided as well
transmit allows you send UCA data directly to the CloudZero API without having to write
code. To use transmit you prepare in advance an ND-JSON formatted input file
which contains one or more correctly formatted JSON UCA records as input.
$ uca transmit --help
Usage: uca transmit [OPTIONS]
Options:
-d, --data TEXT Source data, in text or gzip + text format, supports
file:// or s3:// paths [required]
-o, --output TEXT Instead of sending events to the API, write
events to an output file (note: will overwrite file if it exists)
-t, --transform TEXT Optional transformation script using jq
(https://stedolan.github.io/jq/). Used when the source
data needs modification or cleanup. See README.md for
usage instructions
--help Show this message and exit.
Your input data must be a valid ND-JSON document which can be either a text or gzip file, consisting of one JSON record per line.
For example:
{'timestamp': '2021-03-22 00:00:00+00:00', 'granularity': 'DAILY', 'element-name': 'StateEx', 'filter': {"custom:Environment": ["Production"]}, 'telemetry-stream': 'test-data', 'value': '40.0000'}
{'timestamp': '2021-04-01 00:00:00+00:00', 'granularity': 'DAILY', 'element-name': 'Hooli', 'filter': {"custom:Environment": ["Production"]}, 'telemetry-stream': 'test-data', 'value': '23.0000'}
You can optionally provide a JQ script to transform the source data on the fly before transmission. This is helpful when you have minor or even major changes you want to make to the data quickly, and it would be more complicated or impossible to alter the input data (for example an existing system is producing UCA data in an older format).
The following script requires the id field to be present (records missing this field will be skipped), followed
by setting the target field using a metadata and cost-context field, followed by setting the cost-context to a
constant value and then deleting the metadata and uca fields
select(.id != "")
| .filter = {"tag:environment": [.metadata.environment], "tag:feature": [".metadata.feature"] }
| del(.metadata, .uca)
{
"uca": "v1.5",
"timestamp": "2021-05-25 13:00:00+0000",
"granularity": "HOURLY",
"element-name": "frank",
"telemetry-stream": "test-data",
"value": "1073641",
"metadata": {
"environment": "production",
"feature": "rosebud"
}
}
{
"timestamp": "2021-05-25 13:00:00",
"granularity": "HOURLY",
"id": "frank",
"filter": {
"tag:environment": ["production"],
"tag:feature": ["rosebud"]
},
"telemetry-stream": "Cost-Per-Customer",
"value": "1073641"
}
convert allows you to convert raw data in CSV format to UCA JSON format. To use convert you prepare in advance a CSV file of the
raw unit cost data you wish to convert and then define the mapping of the CSV columns to UCA fields in a configuration file.
$ uca generate --help
Usage: uca convert [OPTIONS]
Options:
-d, --data TEXT Source data, in text or gzip + text format, supports
file:// or s3:// paths [required]
--help Show this message and exit.
{
"version": "1",
"template": {
"timestamp": "$ACCESS_TIME",
"granularity": "HOURLY",
"element-name": "$CUST_ID",
"filter": {
"custom:Environment": [
"Prod"
]
},
"telemetry-stream": "cost-per-customer",
"value": "$DURATION"
},
"settings": {
"api_key": "<YOUR API KEY HERE>" # Also can be provided at runtime via the CLI. Get an API key at https://app.cloudzero.com/organization/api-keys
"convert": {
"stream_name": "cost-per-customer",
"format": "CSV", # currently only CSV is supported
"schema": {
"columns": { # Define one or more columns that will be used from the input CSV file
"CUST_ID": { # The name of the column in the CSV file
"type": "STRING" # Specify the type of the column to ensure proper conversion, can be STRING, NUMBER, DATETIME or TIME
},
"ACCESS_TIME": {
"type": "DATETIME" # DATETIME is a special type that will attempt to convert a wide range of datetime formats (including unix timestamp) to ISO 8601 UTC format
},
"DURATION": {
"type": "TIME", # TIME is a special type that will convert a time duration expressed in HH:mm:ss to seconds
"format": "HH:mm:ss"
}
}
}
}
}
}
Convert raw data from data-file.csv into UCA JSON format using a config-file.json as configuration, and write to output-file.json
$ uca convert -d data-file.csv -c config-file.json -o output-file.json
generate will create UCA data over a defined time period based on rules and source data (CSV) that you define.
Rules enable you to define how the data is to be generated based on the provided source data and can be used to
create static allocations, dynamic allocations, or even random allocations that can be used to verify system operation
or facilitate a demonstration.
$ uca generate --help
Usage: uca generate [OPTIONS]
Options:
-s, --start TEXT start datetime <YYYY-MM-DD HH:MM:SS> [optional]
-e, --end TEXT End datetime <YYYY-MM-DD HH:MM:SS> [optional]
--today Generate events for the current day [optional]
-d, --data TEXT Input UCA data (CSV) [required]
-o, --output TEXT Instead of transmitting the data to the CloudZero API, [optional]
Save the output to a file.
--help Show this message and exit. [optional]
Generate UCA data between 2021-03-13 and 2021-04-07 using data/configuration.json and data/data.csv as input. Performs only a dry run and do not send the results to the CloudZero API
$ uca generate -s 2021-03-13 -e 2021-04-07 -c data/configuration.json -d data/data.csv --dry-run
Generate UCA data using data/configuration.json and data/data.csv (which must contain a timestamp column) as input. Performs only a dry run and do not send the results to the CloudZero API
$ uca generate -s 2021-03-13 -e 2021-04-07 -c data/configuration.json -d data/data.csv --dry-run
The generate command requires a UCA template definition and specific settings that should be applied when generating data. You
can learn more about the UCA format and our UCA Telemetry API here.
{
"version": "1", # can be anything you want that helps you keep track of things
"template": {
"timestamp": "$timestamp", # Timestamp will be replaced automatically or from the input data source
"granularity": "DAILY", # Granularity can be HOURLY, DAILY or MONTHLY. See notes below for usage
"element-name": "$element_name", # Will be replaced using data from your data CSV
"filter": {}, # Use a filter to map spend to a combination of tags, dimensions, accounts, etc...
"telemetry-stream": "test-data", # Unique name for this telemetry stream
"value": "$value" # Will be replaced using generated data or data from your data CSV
},
"settings": {
"generate": {
"mode": "exact", # Can be `exact`, `random`, `jitter` or `allocation`
"jitter": 15, # if mode is jitter, defines the +/- random range to applie to the $value
"allocation": 1000 # if mode is allocation, defines the total amount to be allocated across elements in your data file
"precision": 4 # The number of desired decimal places. If not provided, the default is 4
},
"api_key": "<YOUR API KEY HERE>" # Also can be provided at runtime via the CLI. Get an API key at https://app.cloudzero.com/organization/api-keys
}
}
MONTHLY is a special (non-standard) granularity that can only be used with the UCA toolkit. MONTHLY will expand the input data over the year and month provided to all possible days in a given month to create DAILY UCA telemetry.
For example if you have this input data:
element_name,timestamp,value
"SuperDogs, Inc",2022-10-1,8
"CoolCats, LLC",2022-10-1,229
This will expand to 61 events, 31 for "SuperDogs, Inc" and 31 for "CoolCats, LLC" (31 days in October)
The data CSV defines the input data you wish to feed into the template to produce
matching UCA records for all rows in your data CSV for each hour or day as defined. The following
example has 13 "customers" and can be used as input for all configuration modes (exact, random, jitter, or allocation).
element_name,value,allocation
Sunbank,37,8.5574
SoftwareCorp,17,0.4091
"Parts, Inc.",140,10.9955
Transport Co.,25,6.3033
"WeShipit, Inc.",124,23.7549
CapitalTwo,90,1.5231
Bank of Sokovia,43,0.1198
Makers,9,1.0767
StateEx,40,3.5057
Flitter,15,22.9554
Pets2you,78,6.3358
Hooli,23,4.3294
Massive Dynamic,42,10.1339
Using the exact configuration above, this data will produce UCA events similar to the following:
{'timestamp': '2021-03-22 00:00:00+00:00', 'granularity': 'DAILY', 'element-name': 'StateEx', 'filter': {}, 'telemetry-stream': 'test-data', 'value': '40.0000'}
{'timestamp': '2021-04-01 00:00:00+00:00', 'granularity': 'DAILY', 'element-name': 'Hooli', 'filter': {}, 'telemetry-stream': 'test-data', 'value': '23.0000'}
{'timestamp': '2021-04-01 00:00:00+00:00', 'granularity': 'DAILY', 'element-name': 'Sunbank', 'filter': {}, 'telemetry-stream': 'test-data', 'value': '37.0000'}
{'timestamp': '2021-04-06 00:00:00+00:00', 'granularity': 'DAILY', 'element-name': 'Transport Co.', 'filter': {}, 'telemetry-stream': 'test-data', 'value': '25.0000'}
{'timestamp': '2021-03-19 00:00:00+00:00', 'granularity': 'DAILY', 'element-name': 'StateEx', 'filter': {}, 'telemetry-stream': 'test-data', 'value': '40.0000'}
First create a virtual environment of your choice, and activate it. Then, install UCA Toolkit for local development:
python -m pip install --editable .Run the following commands if you had previously installed it using pip or pipx:
- Uninstall
uca(e.g. pipx uninstall cloudzero-uca-tools) - Create a Virtualenv for your
ucadevelopment - Enter your virtualenv and install all of the necessary dependencies by running make init
- Finally, configure
ucafor use in your environment using pip install -e . - Once that is done, the
ucacli will be available in your shell (you may or may not need to restart your shell).