Looking to gain insights into distribution of stacks throughout a CF
installation? This exporter scrapes the CF API to generate a metric called
cf_stack_count with the label of stack to identify the count of applications
using each stack installed on a CF installation.
It's meant to be run as a cf app, but can easily run wherever you've got Python.
Take a look at manifest.yml: you'll need three bits of information:
CF_API_URL is the full URL for your CF's API (i.e. https://api.sys.cloud.seventhprotocol.com)
CF_USERNAME is an UAA account that has at minimum read-only permissions to all of the /v3/apps endpoint
CF_PASSWORD is the password to the account above.
Optionally, you can also configure:
SCRAPE_INTERVAL which dictates how often the background task of refreshing the metric occurs. By default it's 300 seconds.
LOG_LEVEL which controls verbosity for debugging purposes. By default it's INFO (pretty quiet)
SKIP_SSL_VERIFY for whether or not to skip SSL validation. By default it's False (validates SSL)
INCLUDE_INVALID_STACKS to control whether or not we report stacks that are invalid (i.e typos from developers)
An example CF push:
cf push \
--var cf_api_url=https://api.sys.cloud.seventhprotocol.com \
--var cf_username=readonly \
--var cf_password=deepbluesea
If your CF installation doesn't allow downloading dependencies from the internet (common), run this command to vendor the dependencies prior to pushing:
pip3 download \
--only-binary=:all: \
--python-version=3.12 \
--platform=manylinux2014_x86_64 \
--dest=vendor \
--requirement=requirements.txt
Running a Tanzu Application Service installation, with Healthwatch 2.0? To add the scrape job to HW, go to your OpsMan, then the Healthwatch tile. Navigate to Prometheus > Additional Scrape Jobs.
Paste the following into the scrape job configuration:
job_name: 'stack-usage'
scheme: https
scrape_interval: 60s
static_configs:
- targets: [ 'stack-usage-prometheus-exporter.apps.cloud.seventhprotocol.com' ]
labels:
env: 'staging'
Replace the target with the URL of your exporter instance, and while the label
isn't required, it's recommended to add your environment name to it (i.e.
sandbox, staging, production)