Add Environment & Services support

[MikeGoldsmith]

Which problem is this PR solving?

Adds Environments & Services support to Refinery. This enables users to define environments in their rules configurations, much like a dataset, with custom sampler, sample rate, etc.

Environment names are defined in the rules configuration using the slugified version of the environment name, eg the Environment name "My Cool Company-Production" would become "my_cool_company_production".

Environment names are looked up in the Honeycomb API using the request's API key and then cached for subsequent requests. Environment names are cached for 1 hour.

TODO:

• Add support for configuring environment cache length
• Add tests for existing endpoints with new API keys
• Add tests for EnvironmentCache
• revert the pinned version

Future:

• Review & add default environment to allow env. and dataset with same name in rules file

• Add slugify logic so non-slugified names can be used in rules config, or update Honeycomb API to return un-slugified environment name

• Closes Add support for environment & services #401

Short description of the changes

• Updates Husky to latest (v0.9.0) and update OTLP trace handler to work with batches per service-name
• • This is done once per endpoint, not per event, to prevent unnecessary cache hits
• Update Trace struct to have environment name and GetSamplerKey which returns either the dataset for legacy keys, or environment for non-legacy keys
• When receiving events via OTLP-trace, HTTP /event or /batch, we check the API key and if not legacy (eg 32 chars), lookup environment and add to trace
• Environment lookup is done using the /1/auth Honeycomb API endpoint and result is cached
• Added EnvironmentCache which uses a RWMutex to read/write safety to the underlying map[string]cacheItem where cacheItem holds a TTL and environment name
• Added EnvironmentCacheTTL config option with default of 1 hour "1h"
• Use Trace.GetSamplerKey to resolve rules implementation
• Updates existing tests that expect legacy key to provide a key with 32 characters

[vreynolds]

looks good, just some optimization suggestions. I checked it out locally and did a cursory test with legacy key + dataset-based rules, and new key + environment-based rules

[kentquirk]

My review was mostly limited to looking over the cache stuff.

[MikeGoldsmith]

Need to do a couple of tidy up things before merging

[MikeGoldsmith]

Created follow-up tickets:

• Include refinery version in metrics #406
• Support non-slugified environment names #407
• Add default environment prefix to allow environments and dataset with same name in rules file #408