feat(metrics): add support to persist default dimensions

[heitorlessa]

Issue #, if available: #406

Description of changes:

Checklist

• Meet tenets criteria
• Update tests
• Update docs
• PR title follows conventional commit semantics
• Add new set_default_dimensions method
• Add new clear_default_dimensions method
• Docs: Update testing section to inspect Metrics created
• Docs: Update testing section to reset Default Dimensions
• Docs: Banner about adding metrics outside the handler
• Docs: Banner about sparse cold start metrics for cost
• Docs: Add Functional testing section

UX

Persists two dimensions across all metrics, every Metrics instance throughout the codebase, and across all Lambda invocations.

from aws_lambda_powertools import Metrics
from aws_lambda_powertools.metrics import MetricUnit

metrics = Metrics(namespace="ExampleApplication", service="booking")
metrics.set_default_dimensions(environment="prod", another="one")

@metrics.log_metrics
def lambda_handler(evt, ctx):
  metrics.add_metric(name="SuccessfulBooking", unit=MetricUnit.Count, value=1)

Breaking change checklist

RFC issue #:

• Migration process documented
• Implement warnings (if it can live side by side)

By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.

[pcolazurdo]

Only a few documentation comments and a question around the behavior of clear_metrics()

[pcolazurdo]

I'd like this approach for some use cases but I think it may add a "burden" on the dev UX to remember where the defaults where set. I'd prefer an approach where the decorator has something like:

@metrics.log_metrics(default_dimensions=**dimensions)

Which keeps the defaults very close to the function. I wouldn't expect this defaults to be shared across other decorators in my opinion - i.e. if the code is reusing a library that has it's own defaults

[heitorlessa]

I'd like this approach for some use cases but I think it may add a "burden" on the dev UX to remember where the defaults where set. I'd prefer an approach where the decorator has something like:

@metrics.log_metrics(default_dimensions=**dimensions)

Which keeps the defaults very close to the function. I wouldn't expect this defaults to be shared across other decorators in my opinion - i.e. if the code is reusing a library that has it's own defaults

Yeah, I'd have loved to go down that route too, however param= doesn't accept unpacking a dictionary, it'd have to be @metrics.log_metrics(**dimensions) in which case makes it less explicit to whoever reading the code what this does.

I also experimented with:

dimensions = [{"name": "dimension_1"}, "value": "value"}, {"name": "dimension_2"}, "value": "value"}]
@metrics.log_metrics(default_dimensions=dimensions)

But this was more explicit and can work outside Lambda handler like in Lambda layers code, other modules, etc.

metrics = Metrics()

metrics.set_default_dimensions(environment="prod", dimension_one="value", ...)

[pcolazurdo]

Wouldn't something like this below help?

@my_metrics.log_metrics(default_dimensions={"environment": "dev"})
def lambda_handler(evt, context):
  return True

I have a sample version here: https://github.com/heitorlessa/aws-lambda-powertools-python/compare/feat/metrics-default-dimensions...pcolazurdo:pr_410?expand=1

[heitorlessa]

This does @pcolazurdo - I was initially thinking the name=, value=, which didn't fit.

I'll include that in today now that the implementation is already done

[pcolazurdo]

Looks great

[codecov-commenter]

Codecov Report

Merging #410 (a1da396) into develop (68d4110) will increase coverage by 0.00%.
The diff coverage is 100.00%.

@@           Coverage Diff            @@
##           develop     #410   +/-   ##
========================================
  Coverage    99.89%   99.89%           
========================================
  Files          100      100           
  Lines         3779     3827   +48     
  Branches       175      178    +3     
========================================
+ Hits          3775     3823   +48     
  Misses           1        1           
  Partials         3        3           

Impacted Files	Coverage Δ
aws_lambda_powertools/metrics/metrics.py	100.00% <100.00%> (ø)
aws_lambda_powertools/logging/logger.py	100.00% <0.00%> (ø)
aws_lambda_powertools/logging/formatter.py	100.00% <0.00%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 68d4110...a1da396. Read the comment docs.