duckula.handler — nomnom/duckula 0.8.0

Default Duckula handler. It talks JSON but can validate requests with provided Avro schemas

build clj

(build config)

Sort of a router, but does validation. Config has to have the form of: { "/some/route" {:handler ring.rquest.handler/function :response "path/to/avro/response/Schema" :request "path/to/avro/request/Schema"} ; .. and more "/no/validation" {:handler ring.req.handler/no-validation} } Handler assumes that requests use application/json for input and output as content types. For 100% avro input/input we'd need a slightly different builder (and ring middlewares). It depends on a component implementing duckula.prococol/Monitoring protocol

• request count (success, error, failure)
• request timing
• track exceptions

build-metric-keys clj

(build-metric-keys {:keys [endpoints prefix] :as config})

For each endpoint it constructs a list of metric keys to be used when tracking timing, rates and errors Example for name test-api and /some/endpoint it will create test-api.some.endpoint test-api.some.endpoint.success test-api.some.endpoint.error test-api.some.endpoint.failure

build-route-map clj

(build-route-map {:keys [prefix endpoints] :as config})

Turns static config (documented below) into a map of function maps:

• request handler
• request input validator
• request output validator Validators use Avro to ensure passed in data is ok

not-found-404 clj

(not-found-404 & _)

not-found-metrics clj

use-kebab-case? clj

(use-kebab-case? {:keys [mangle-names? kebab-case-names?]})

validate-with-tag clj

(validate-with-tag tag validator-fn input monitoring)

Runs validation function and re-throws the exception with extra info attached. Each generated avro validator function carries metadata with schema name