Liking cljdoc? Tell your friends :D

Getting Started

cartus is a structured logging abstraction for logging data rich events with support for multiple logging backends, currently including:

  • a test logger for collecting logs in memory and asserting against them;
  • a cambium logger for logging out via SLF4J and logback; and
  • a null logger that ignores all logged events, useful as a default implementation or to turn off logging completely.

cartus is heavily inspired by JUXT's blog post on logging which recommends treating the logger as an injectable dependency like any other. This brings a number of benefits:

  • Testing for log events becomes easy, since the logger implementation can be switched out for one more amenable to testing.
  • More generally, different logging backends can be supported in different contexts.
  • Since the logger is passed as an explicit dependency, it can be transformed in various ways, such as by adding context to all log events or filtering log events to only those of a certain set of types within a particular scope.

Contents

Installation

Add the following to your project.clj file:

[io.logicblocks/cartus.core "0.1.16"]

Configuring a backend

The cartus.test/logger backend

The [[cartus.test/logger]] backend captures all logged events in memory in an atom, allowing your tests to assert that log events took place.

To install the cartus.test/logger backend, add the following to your project.clj file:

[io.logicblocks/cartus.test "0.1.16"]

To create a cartus.test/logger:

(require '[cartus.test])

(def logger (cartus.test/logger))

The cartus.cambium/logger backend

The [[cartus.cambium/logger]] backend passes all logged events to cambium which in turn uses SLF4J and logback to log out the log event either in a plain text format or as JSON.

To install the cartus.cambium/logger backend, add the following to your project.clj file:

[io.logicblocks/cartus.cambium "0.1.16"]

Additionally, you must choose a codec and backend for cambium, which has implementations for plain text logging and for JSON logging.

To use plain text logs, add the following to your project.clj file:

[cambium/cambium.codec-simple "0.9.3"]

and add logback configuration (typically at resources/logback.xml) containing something like the following:

<configuration>
  <appender name="STDOUT" class="ch.qos.logback.core.ConsoleAppender">
    <encoder>
      <pattern>
        %d{HH:mm:ss.SSS} [%thread] %-5level %logger{36} - %msg { %mdc }%n
      </pattern>
    </encoder>
  </appender>
  <root level="debug">
    <appender-ref ref="STDOUT" />
  </root>
</configuration>

To use JSON logs, add the following to your project.clj file:

[cambium/cambium.codec-cheshire "0.9.3"]
[cambium/cambium.logback.json   "0.4.3"]

and add logback configuration (again, typically at resources/logback.xml) containing something like the following:

<configuration>
  <appender name="STDOUT" class="ch.qos.logback.core.ConsoleAppender">
    <encoder class="ch.qos.logback.core.encoder.LayoutWrappingEncoder">
      <layout class="cambium.logback.json.FlatJsonLayout">
        <jsonFormatter class="ch.qos.logback.contrib.jackson.JacksonJsonFormatter">
          <prettyPrint>true</prettyPrint>
        </jsonFormatter>
        <timestampFormat>yyyy-MM-dd'T'HH:mm:ss.SSS'Z'</timestampFormat>
        <timestampFormatTimezoneId>UTC</timestampFormatTimezoneId>
        <appendLineSeparator>true</appendLineSeparator>
      </layout>
    </encoder>
  </appender>
  <root level="debug">
    <appender-ref ref="STDOUT" />
  </root>
</configuration>

For more details of configuring cambium backends, see the cambium documentation.

Once a cambium codec and backend has been configured, cambium must be initialised as the system starts up. This can be achieved with the following:

(require '[cartus.cambium])

(cartus.cambium/initialise)

See [[cartus.cambium/initialise]] for more details on initialisation options.

Once cambium is initialised, to create a cartus.cambium/logger:

(def logger (cartus.cambium/logger))

The cartus.null/logger backend

The [[cartus.null/logger]] backend ignores all logged events which can be useful to turn off logging completely or provide a default implementation in the case that one is not provided.

To install the cartus.null/logger backend, add the following to your project.clj file:

[io.logicblocks/cartus.null "0.1.16"]

To create a cartus.null/logger:

(require '[cartus.null])

(def logger (cartus.null/logger))

Custom backends

All cartus.core functions expect an implementation of the [[cartus.core/Logger]] protocol.

For example, to create a naive StdoutLogger which prints all log events to the standard output stream as EDN:

(ns example.stdout
  (:require
   [cartus.core :as core]))

(defrecord StdoutLogger
  []
  core/Logger
  (log [_ level type context opts]
    (println 
      (merge {:level level
            :type type
            :context context}
        opts))))

(defn logger []
  (map->StdoutLogger {}))

Logging events

Log events:

  • are logged at a specific level, typically one of :trace, :debug, :info, :warn, :error or :fatal;
  • have a type, expressed as a keyword, such as :some.namespace/some.event.type;
  • include a map of additional context, which can be deeply nested;
  • optionally include a message giving a textual description of the event;
  • optionally include an exception including more details of an error;
  • optionally include a meta map, including any metadata at the point of the call to the logger such as the line, column and namespace of the call.

To log events at each respective level, you can use the convenience macros [[cartus.core/trace]], [[cartus.core/debug]], [[cartus.core/info]], [[cartus.core/warn]], [[cartus.core/error]] and [[cartus.core/fatal]], which also capture metadata for the call site automatically:

(ns example.logging
  (:require
   [cartus.core :as log]
   [cartus.cambium]))

(cartus.cambium/initialise)

(def logger (cartus.cambium/logger))

(log/trace logger ::system.started 
  {:flags {:in-memory-database false
           :live-services false}})
; => 
; "{
;   \"timestamp\" : \"2020-07-18T16:59:36.625Z\",
;   \"level\" : \"TRACE\",
;   \"thread\" : \"nREPL-session-ea7676d2-b76c-44c4-bc2b-d1c2f8515c0b\",
;   \"ns\" : \"example.logging\",
;   \"line\" : 10,
;   \"column\" : 1,
;   \"flags\" : {
;     \"in-memory-database\" : false,
;     \"live-services\" : false
;   },
;   \"type\" : \"example.logging/system.started\",
;   \"logger\" : \"example.logging\",
;   \"message\" : \"example.logging/system.started\",
;   \"context\" : \"default\"
; }
; "

(log/info logger ::database.migrating
  {:target-schema-version 10})
; => 
; "{
;   \"timestamp\" : \"2020-07-18T17:00:54.954Z\",
;   \"level\" : \"INFO\",
;   \"thread\" : \"nREPL-session-ea7676d2-b76c-44c4-bc2b-d1c2f8515c0b\",
;   \"ns\" : \"example.logging\",
;   \"line\" : 32,
;   \"column\" : 1,
;   \"target-schema-version\" : 10,
;   \"type\" : \"example.logging/database.migrating\",
;   \"logger\" : \"example.logging\",
;   \"message\" : \"example.logging/database.migrating\",
;   \"context\" : \"default\"
; }
; "

(log/debug logger ::database.migrated
  {:current-schema-version 10 :elapsed-millis 568}
  {:message "Success!"})
; =>
; "{
;   \"timestamp\" : \"2020-07-18T17:02:10.798Z\",
;   \"level\" : \"DEBUG\",
;   \"thread\" : \"nREPL-session-ea7676d2-b76c-44c4-bc2b-d1c2f8515c0b\",
;   \"ns\" : \"example.logging\",
;   \"line\" : 50,
;   \"current-schema-version\" : 10,
;   \"column\" : 1,
;   \"elapsed-millis\" : 568,
;   \"type\" : \"example.logging/database.migrated\",
;   \"logger\" : \"example.logging\",
;   \"message\" : \"Success!\",
;   \"context\" : \"default\"
; }

(log/error logger ::cache.connecting
  {:address "tcp://cache.example.com:3456"}
  {:exception (ex-info "Connection failed" {:timeout-millis 500})})
; =>
; "{
;   \"timestamp\" : \"2020-07-18T17:03:13.756Z\",
;   \"level\" : \"ERROR\",
;   \"thread\" : \"nREPL-session-ea7676d2-b76c-44c4-bc2b-d1c2f8515c0b\",
;   \"address\" : \"tcp://cache.example.com:3456\",
;   \"ns\" : \"example.logging\",
;   \"line\" : 69,
;   \"column\" : 1,
;   \"type\" : \"example.logging/cache.connecting\",
;   \"logger\" : \"example.logging\",
;   \"message\" : \"example.logging/cache.connecting\",
;   \"context\" : \"default\",
;   \"exception\" : \"clojure.lang.ExceptionInfo: Connection failed\
; \\tat example.logging$eval6273.invokeStatic(form-init17428929335843016668.clj:3)\
; \\tat example.logging$eval6273.invoke(form-init17428929335843016668.clj:1)\
; \\tat clojure.lang.Compiler.eval(Compiler.java:7177)\
; \\tat clojure.lang.Compiler.eval(Compiler.java:7132)\
; \\tat clojure.core$eval.invokeStatic(core.clj:3214)\
; \\tat clojure.core$eval.invoke(core.clj:3210)\
; \\tat nrepl.middleware.interruptible_eval$evaluate$fn__3914.invoke(interruptible_eval.clj:91)\
; \\tat clojure.main$repl$read_eval_print__9086$fn__9089.invoke(main.clj:437)\
; \\tat clojure.main$repl$read_eval_print__9086.invoke(main.clj:437)\
; \\tat clojure.main$repl$fn__9095.invoke(main.clj:458)\
; \\tat clojure.main$repl.invokeStatic(main.clj:458)\
; \\tat clojure.main$repl.doInvoke(main.clj:368)\
; \\tat clojure.lang.RestFn.invoke(RestFn.java:1523)\
; \\tat nrepl.middleware.interruptible_eval$evaluate.invokeStatic(interruptible_eval.clj:84)\
; \\tat nrepl.middleware.interruptible_eval$evaluate.invoke(interruptible_eval.clj:56)\
; \\tat nrepl.middleware.interruptible_eval$interruptible_eval$fn__3940$fn__3944.invoke(interruptible_eval.clj:155)\
; \\tat clojure.lang.AFn.run(AFn.java:22)\
; \\tat nrepl.middleware.session$session_exec$main_loop__4041$fn__4045.invoke(session.clj:190)\
; \\tat nrepl.middleware.session$session_exec$main_loop__4041.invoke(session.clj:189)\
; \\tat clojure.lang.AFn.run(AFn.java:22)\
; \\tat java.lang.Thread.run(Thread.java:748)\
; \"
; }
; "

(log/fatal logger ::exception.unknown)
; =>
; "{
;   \"timestamp\" : \"2020-07-18T17:06:22.932Z\",
;   \"level\" : \"ERROR\",
;   \"thread\" : \"nREPL-session-f070aa75-a886-4207-83f8-65f73c0ba556\",
;   \"ns\" : \"example.logging\",
;   \"line\" : 50,
;   \"column\" : 1,
;   \"type\" : \"example.logging/exception.unknown\",
;   \"logger\" : \"example.logging\",
;   \"context\" : \"default\"
; }

Alternatively, you can use the cartus.core/log function from [[cartus.core/Logger]] to log directly, giving full control over all aspects of the log event:

(ns example.logging
  (:require
   [cartus.core :as log]
   [cartus.cambium]))

(cartus.cambium/initialise)

(def logger (cartus.cambium/logger))

(log/log logger :info ::system.started
  {:flags {:in-memory-database false
             :live-services false}}
  {:meta {:line 10 :column 20 :ns (find-ns 'example.other)}})
; =>
; "{
;   \"timestamp\" : \"2020-07-18T17:35:50.896Z\",
;   \"level\" : \"INFO\",
;   \"thread\" : \"nREPL-session-ea7676d2-b76c-44c4-bc2b-d1c2f8515c0b\",
;   \"ns\" : \"example.other\",
;   \"line\" : 10,
;   \"column\" : 20,
;   \"flags\" : {
;     \"in-memory-database\" : false,
;     \"live-services\" : false
;   },
;   \"type\" : \"example.logging/system.started\",
;   \"logger\" : \"example.other\",
;   \"message\" : \"example.logging/system.started\",
;   \"context\" : \"default\"
; }
; "

Note that in the case you use cartus.core/log directly, no metadata is captured so metadata must be provided explicitly.

Applying transformations to loggers

Since cartus.core functions accept the logger as an explicit dependency, you can easily apply transformations to the logger as it is passed down through function calls. This allows additional behaviour to be added to a logger in a scoped manner, irrespective of backend used.

Setting logger context

To add logger context to a logger instance, use [[cartus.core/with-context]]. This function returns a new logger which will merge the provided context map with the context map provided at log time, with the log time context taking preference:

(ns example.logging
  (:require
   [cartus.core :as log]
   [cartus.cambium]))

(cartus.cambium/initialise)

(def standard-logger (cartus.cambium/logger))
(def contextual-logger (cartus.core/with-context standard-logger
                        {:request-id 5 :user-id 15}))

(log/info contextual-logger ::service.requesting
  {:request-id 10 :endpoint-name "check"})
; =>
; "{
;   \"timestamp\" : \"2020-07-18T17:51:35.862Z\",
;   \"level\" : \"INFO\",
;   \"thread\" : \"nREPL-session-ea7676d2-b76c-44c4-bc2b-d1c2f8515c0b\",
;   \"endpoint-name\" : \"check\",
;   \"ns\" : \"example.logging\",
;   \"line\" : 12,
;   \"user-id\" : 15,
;   \"column\" : 1,
;   \"request-id\" : 10,
;   \"type\" : \"example.logging/service.requesting\",
;   \"logger\" : \"example.logging\",
;   \"message\" : \"example.logging/service.requesting\",
;   \"context\" : \"default\"
; }
; "

Filtering by log levels

To retain log events based on their level, use [[cartus.core/with-levels-retained]]. This function returns a new logger which will drop any log events that do not match the specified criteria based on a couple of different variants as discussed below.

To retain log events having a level within a provided set of levels:

(ns example.logging
  (:require
   [cartus.core :as log]
   [cartus.cambium]))

(cartus.cambium/initialise)

(def standard-logger (cartus.cambium/logger))
(def filtered-logger (cartus.core/with-levels-retained standard-logger
                        #{:warn :error}))

(log/info filtered-logger ::service.requesting
  {:request-id 10 :endpoint-name "check"})
; => nothing logged

(log/warn filtered-logger ::service.slow
  {:latency-millis 5486})
; "{
;   \"timestamp\" : \"2020-07-19T15:21:42.436Z\",
;   \"level\" : \"WARN\",
;   \"thread\" : \"nREPL-session-796b74ce-6705-4cf7-a55c-d95b089b8b34\",
;   \"latency-millis\" : 5486,
;   \"ns\" : \"example.logging\",
;   \"line\" : 15,
;   \"column\" : 1,
;   \"type\" : \"example.logging/service.slow\",
;   \"logger\" : \"example.logging\",
;   \"message\" : \"example.logging/service.slow\",
;   \"context\" : \"default\"
; }
; "

To retain log events having a level greater than or equal to a specified level in severity:

(ns example.logging
  (:require
   [cartus.core :as log]
   [cartus.cambium]))

(cartus.cambium/initialise)

(def standard-logger (cartus.cambium/logger))
(def filtered-logger 
  (cartus.core/with-levels-retained standard-logger >= :info))

(log/debug filtered-logger ::database.connection-pool.requesting
  {:timeout-millis 200})
; => nothing logged

(log/info filtered-logger ::database.querying
  {:query-name :find-user})
; "{
;   \"timestamp\" : \"2020-07-19T15:28:23.843Z\",
;   \"level\" : \"INFO\",
;   \"thread\" : \"nREPL-session-796b74ce-6705-4cf7-a55c-d95b089b8b34\",
;   \"query-name\" : "find-user",
;   \"ns\" : \"example.logging\",
;   \"line\" : 15,
;   \"column\" : 1,
;   \"type\" : \"example.logging/database.querying\",
;   \"logger\" : \"example.logging\",
;   \"message\" : \"example.logging/database.querying\",
;   \"context\" : \"default\"
; }
; "

The arity-3 version of [[cartus.core/with-levels-retained]] accepts the operators >=, >, =, < and <=.

To ignore log events based on their level, use [[cartus.core/with-levels-ignored]]. This function returns a new logger which will drop any log events that match the specified criteria based on a couple of different variants as discussed below.

To ignore log events having a level within a provided set of levels:

(ns example.logging
  (:require
   [cartus.core :as log]
   [cartus.cambium]))

(cartus.cambium/initialise)

(def standard-logger (cartus.cambium/logger))
(def filtered-logger (cartus.core/with-levels-ignored standard-logger
                        #{:warn :error}))

(log/info filtered-logger ::service.requesting
  {:request-id 10 :endpoint-name "check"})
; =>
; "{
;   \"timestamp\" : \"2020-07-19T17:35:12.894Z\",
;   \"level\" : \"INFO\",
;   \"thread\" : \"nREPL-session-d8857a81-72f5-41ce-a107-eaecc7b75805\",
;   \"request-id\" : 10,
;   \"endpoint-name\" : \"check\",
;   \"ns\" : \"example.logging\",
;   \"line\" : 12,
;   \"column\" : 1,
;   \"type\" : \"example.logging/service.requesting\",
;   \"logger\" : \"example.logging\",
;   \"message\" : \"example.logging/service.requesting\",
;   \"context\" : \"default\"
; }
; "

(log/warn filtered-logger ::service.slow
  {:latency-millis 5486})
; => nothing logged

To ignore log events having a level less than or equal to a specified level in severity:

(ns example.logging
  (:require
   [cartus.core :as log]
   [cartus.cambium]))

(cartus.cambium/initialise)

(def standard-logger (cartus.cambium/logger))
(def filtered-logger 
  (cartus.core/with-levels-ignored standard-logger < :info))

(log/debug filtered-logger ::database.connection-pool.requesting
  {:timeout-millis 200})
; => nothing logged

(log/info filtered-logger ::database.querying
  {:query-name :find-user})
; "{
;   \"timestamp\" : \"2020-07-19T17:48:13.983Z\",
;   \"level\" : \"INFO\",
;   \"thread\" : \"nREPL-session-b8bbb4af-f967-49eb-8ef3-9ad8f93376e8\",
;   \"query-name\" : "find-user",
;   \"ns\" : \"example.logging\",
;   \"line\" : 15,
;   \"column\" : 1,
;   \"type\" : \"example.logging/database.querying\",
;   \"logger\" : \"example.logging\",
;   \"message\" : \"example.logging/database.querying\",
;   \"context\" : \"default\"
; }
; "

The arity-3 version of [[cartus.core/with-levels-ignored]] accepts the operators >=, >, =, < and <=.

Filtering by log types

To retain log events based on their type, use [[cartus.core/with-types-retained]]. This function returns a new logger which will drop any log events not having a type in the provided set of types:

(ns example.logging
  (:require
   [cartus.core :as log]
   [cartus.cambium]))

(cartus.cambium/initialise)

(def standard-logger (cartus.cambium/logger))
(def filtered-logger (cartus.core/with-types-retained standard-logger
                        #{::order.rejected ::order.approved}))

(log/debug filtered-logger ::order.pending
  {:outstanding-steps [:payment]})
; => nothing logged

(log/info filtered-logger ::order.rejected
  {:reason :card-payment-failed})
; "{
;   \"timestamp\" : \"2020-07-19T18:56:22.674Z\",
;   \"level\" : \"INFO\",
;   \"thread\" : \"nREPL-session-5d899bb7-3d16-4d84-81e0-691fc2df2c66\",
;   \"reason\" : "card-payment-failed",
;   \"ns\" : \"example.logging\",
;   \"line\" : 15,
;   \"column\" : 1,
;   \"type\" : \"example.logging/order.rejected\",
;   \"logger\" : \"example.logging\",
;   \"message\" : \"example.logging/order.rejected\",
;   \"context\" : \"default\"
; }
; "

To ignore log events based on their type, use [[cartus.core/with-types-ignored]]. This function returns a new logger which will drop any log events having a type in the provided set of types:

(ns example.logging
  (:require
   [cartus.core :as log]
   [cartus.cambium]))

(cartus.cambium/initialise)

(def standard-logger (cartus.cambium/logger))
(def filtered-logger (cartus.core/with-types-ignored standard-logger
                        #{::order.pending}))

(log/debug filtered-logger ::order.pending
  {:outstanding-steps [:payment]})
; => nothing logged

(log/info filtered-logger ::order.rejected
  {:reason :card-payment-failed})
; "{
;   \"timestamp\" : \"2020-07-19T18:57:23.682Z\",
;   \"level\" : \"INFO\",
;   \"thread\" : \"nREPL-session-5d899bb7-3d16-4d84-81e0-691fc2df2c66\",
;   \"reason\" : "card-payment-failed",
;   \"ns\" : \"example.logging\",
;   \"line\" : 15,
;   \"column\" : 1,
;   \"type\" : \"example.logging/order.rejected\",
;   \"logger\" : \"example.logging\",
;   \"message\" : \"example.logging/order.rejected\",
;   \"context\" : \"default\"
; }
; "

Applying arbitrary transformations

More generally, a logger can be seen as a stream of log events. With this in mind, cartus supports transducers being applied to loggers. This allows arbitrarily complex transformations to be built up using Clojure's existing transducer support.

For example, to remove events that contain sensitive information and then map all context keys to snake case before logging an event, you can use the following:

(ns example.logging
  (:require
   [clojure.string :as string]
   [clojure.walk :as walk]

   [cartus.core :as log]
   [cartus.cambium]))

(cartus.cambium/initialise)

(defn transform-keys
  [t coll]
  (let [f (fn [[k v]] [(t k) v])]
    (walk/postwalk (fn [x] (if (map? x) (into {} (map f x)) x)) coll)))

(defn ->camel-case [k]
  (keyword 
    (string/replace (name k) #"-(\w)"
      #(string/upper-case (second %1)))))

(def standard-logger (cartus.cambium/logger))
(def filtered-logger 
  (cartus.core/with-transformation standard-logger
    (comp 
      (filter #(not (contains? (:context %) :password)))
      (map #(assoc % :context (transform-keys ->camel-case (:context %)))))))

(log/debug filtered-logger ::order.pending
  {:outstanding-steps [:payment]
   :line-items [:item-1 :item-2]})
; "{
;   \"timestamp\" : \"2020-07-23T08:51:14.165Z\",
;   \"level\" : \"DEBUG\",
;   \"thread\" : \"nREPL-session-86b79ffc-6ce9-43af-91a9-0b1de812e86c\",
;   \"lineItems\" : [\"item-1\",\"item-2\"],
;   \"ns\" : \"example.logging\",
;   \"line\" : 1,
;   \"column\" : 1,
;   \"outstandingSteps\" : [\"payment\"],
;   \"type\" : \"example.logging/order.pending\",
;   \"logger\" : \"example.logging\",
;   \"message\" : \"example.logging/order.pending\",
;   \"context\" : \"default\"
; }
; "

(log/info filtered-logger ::registration.in-progress
  {:password "super-secret"})
; => nothing logged

All other transformations in [[cartus.core]] use the transducer support under the covers and can be used as examples when implementing your own.

Testing for log events

With the help of the [[cartus.test/logger]] and utilities in the [[cartus.test]], namespace you can assert that log events occurred from within your tests.

Asserting directly against logged events

(ns example.subject
  (:require
   [cartus.core :as log]))

(defn +-with-logging [logger & vals]
  (log/info logger ::summing.values {:values vals})
  (apply + vals))

(ns example.subject-test
  (:require
   [clojure.test :refer :all]

   [cartus.test :as log-test]
   
   [example.subject :as subject]))

(deftest logs-while-summing
  (let [logger (log-test/logger)]
    (subject/+-with-logging logger 1 2 3)
    (is (= [{:level   :info
             :type    :example.subject/summing.values
             :context {:values [1 2 3]}
             :meta    {:ns (find-ns 'example.subject)
                       :line 6
                       :column 3}}]
      (log-test/events logger)))))
; => passes

Using the logged? assertion

Asserting directly against log events returned by [[cartus.test/events]] can be cumbersome. To make log assertions more intentional, use the [[cartus.test/logged?]] assertion, which implements the clojure.test/assert-expr multimethod, extending clojure.test/is to respond to a logged? form:

(deftest logs-while-summing
  (let [logger (cartus.test/logger)]
    (subject/+-with-logging logger 1 2 3)
    (is (logged? logger 
          {:level   :info
           :type    :example.subject/summing.values
           :context {:values [1 2 3]}
           :meta    {:ns (find-ns 'example.subject)
                     :line 6
                     :column 3}}))))
; => passes

Using modifiers with logged?

By default, logged? matches fuzzily against the events logged to the logger. As a result, log specs don't need to include all attributes and can instead include a subset:

(deftest logs-while-summing
  (let [logger (cartus.test/logger)]
    (subject/+-with-logging logger 1 2 3)
    (is (logged? logger 
          {:level   :info
           :type    :example.subject/summing.values
           :context {:values [1 2 3]}}))))
; => passes

This is useful when the metadata of the logged event is unimportant or where you don't care about some part of the logged event, say, the level at which the event was logged.

If instead you want to match strictly against the events logged to the logger, pass the :strict-contents modifier:

(deftest logs-while-summing
  (let [logger (cartus.test/logger)]
    (subject/+-with-logging logger 1 2 3)
    (is (logged? logger #{:strict-contents}
          {:level   :info
           :type    :example.subject/summing.values
           :context {:values [1 2 3]}}))))
; => fails, due to missing `:meta` attribute

The logged? assertion accepts a variable number of log specs, to assert against a sequence of log events. By default, logged? expects the order of log events to match the log specs but allows surplus log events in the log stream and allows log events to occur between those events specified by the provided log specs. As such, it is enough for matching log events to appear in the correct order somewhere in the log stream:

(deftest logs-while-summing
  (let [logger (cartus.test/logger)]
    
    (subject/+-with-logging logger 1 2 3)
    (subject/+-with-logging logger 4 5 6)
    (subject/+-with-logging logger 7 8 9)
    (subject/+-with-logging logger 10 11 12)
    (subject/+-with-logging logger 13 14 15)

    (is (logged? logger 
          {:type    :example.subject/summing.values
           :context {:values [4 5 6]}}
          {:type    :example.subject/summing.values
           :context {:values [10 11 12]}}))))
; => passes

If instead, you want to assert that a set of log events occur, but the order is not important, pass the :in-any-order modifier:

(deftest logs-while-summing
  (let [logger (cartus.test/logger)]
    
    (subject/+-with-logging logger 1 2 3)
    (subject/+-with-logging logger 4 5 6)
    (subject/+-with-logging logger 7 8 9)
    (subject/+-with-logging logger 10 11 12)
    (subject/+-with-logging logger 13 14 15)

    (is (logged? logger #{:in-any-order} 
          {:type    :example.subject/summing.values
           :context {:values [10 11 12]}}
          {:type    :example.subject/summing.values
           :context {:values [4 5 6]}}))))
; => passes

This can be useful when concurrent processes are logging to the same logger since event ordering in such a case is not guaranteed.

Sometimes, you might want to assert that a strict sequence of log events has occurred and don't want to tolerate surplus log events. To assert that only the specified log events have occurred, use the :only modifier:

(deftest logs-while-summing
  (let [logger (cartus.test/logger)]
    
    (subject/+-with-logging logger 1 2 3)
    (subject/+-with-logging logger 4 5 6)
    (subject/+-with-logging logger 7 8 9)
    (subject/+-with-logging logger 10 11 12)
    (subject/+-with-logging logger 13 14 15)

    (is (logged? logger #{:only} 
          {:type    :example.subject/summing.values
           :context {:values [4 5 6]}}
          {:type    :example.subject/summing.values
           :context {:values [10 11 12]}}))))
; => fails, since there are many surplus log events

Again, the default behaviour is that log events are expected in the order in which they are specified. If instead you want to assert that only the specified log events have occurred but where order is not important, use both the :only and :in-any-order modifiers:

(deftest logs-while-summing
  (let [logger (cartus.test/logger)]
    
    (subject/+-with-logging logger 1 2 3)
    (subject/+-with-logging logger 4 5 6)
    
    (is (logged? logger #{:only :in-any-order} 
          {:type    :example.subject/summing.values
           :context {:values [4 5 6]}}
          {:type    :example.subject/summing.values
           :context {:values [1 2 3]}}))))
; => passes

The available modifiers are:

  • #{:in-order :in-any-order} defining ordering constraints
  • #{:only :at-least} defining cardinality constraints
  • #{:fuzzy-contents :strict-contents} defining content constraints

and any combination of one from each set is permitted. The effective default modifiers are: #{:in-order :at-least :fuzzy-contents}.

Advanced usage of logged?

Internally, the logged? assertion uses the matcher-combinators library which allows for more advanced matching. Specifically, matcher-combinators supports predicate matchers as well as more targetted collection matching.

For example, to ensure that a log event occurs with one of a number of types, you can do the following:

(is (logged? logger
      {:level :info
       :type #(or (= :some.ns/type-1 %) (= :some.ns/type-2 %))}))

This will assert that at least a log event with a type of :some.ns/type-1 or :some.ns/type-2 occurs at log level :info.

Predicates can occur deep inside a nested log spec, for example:

(is (logged? logger
      {:level :info
       :type  :some.ns/type-1
       :context {:factors [#(> % 3) #(> % 5) 8]}}))

In addition to predicate matchers, matcher-combinators allows you to more tightly or loosely control the make up of collection types. By default, maps are matched using a matcher called embeds which allows surplus attributes to be present. If you wish to exactly match a map in the log event context, whilst retaining fuzzy matching on the log event overall you can do the following:

(require '[matcher-combinators.matchers :as mc-matchers])

(is (logged? logger
      {:level :info
       :type  :some.ns/type-1
       :context {:reference 
                 (mc-matchers/equals 
                   {:id "123"
                    :region "us"
                    :account-id "456"})}}))

Internally, the logged? modifiers are implemented as combinations of different matcher-combinator matchers, allowing for more flexible log specs than provided by the modifiers alone.

For more details on how to compose more complex matchers, see the matcher-combinator documentation.

Can you improve this documentation? These fine people already did:
Circle CI & Toby Clemson
Edit on GitHub

cljdoc is a website building & hosting documentation for Clojure/Script libraries

× close