Berlin 2013 - Riemann Workshop - Pierre-Yves Ritschard

Transcript

1. @pyr
   

   View Slide

2. https://github.com/pyr
   

   View Slide

3. ● Riemann concepts
   ● In-app events
   ● System metrics
   ● Just enough clojure
   ● Common use cases
   Agenda
   

   View Slide

4. Event stream processor
   ● Process
   ● Analyze
   ● React
   

   View Slide

5. View Slide

6. {:host “foo1”
   :service “cpu %”
   :metric 50.85
   :state “ok”
   :description “System CPU gauge”}
   

   View Slide

7. Where do events come from
   ● Your app
   ● Collectd
   ● Riemann-health
   ● Logstash
   

   View Slide

8. import bernhard
   c = bernhard.Client()
   c.send({‘host’: “appserver”,
   ‘service’: “sql query time”,
   ‘metric’: 200})
   

   View Slide

9. require ‘riemann/client’
   c = Riemann::Client.new
   c << {host: “appserver”,
   service: “sql query time”,
   metric: 200}
   

   View Slide

10. LoadPlugin write_riemann
    
    
    Host “riemann-host”
    
    
    

    View Slide

11. Event lifecycle
    ● Enters riemann
    ● Follows streams
    ● Optionally indexed
    

    View Slide

12. Looking at events
    ● Queries
    ● Dashboard
    

    View Slide

13. require ‘riemann/client’
    c = Riemann::Client.new
    c[‘true’]
    c[‘service = “cpu”’]
    c[‘tagged “web-servers”’]
    

    View Slide

14. Configuration API Features
    ● Percentiles
    ● EWMA
    ● Fixed / Moving event windows
    ● Event coalescing
    ● Many more
    

    View Slide

15. Configuration
    ● Configuration is code
    ● Code is in clojure
    

    View Slide

16. Clojure
    

    View Slide

17. Clojure
    ● Functional Lisp
    ● Immutable
    ● Data first, not objects first
    ● Lazy
    ● Concurrent
    

    View Slide

18. Clojure
    (defn hello [name]
    (println “hello” name)
    (def a-map {:some-key “some value”})
    (def an-array [0 1 2 :foo])
    (def a-set #{:a :b :c})
    (def a-pattern #”^(.*)[ \t]+=(.*)[ \t]+$”)
    (def inline-fn #(+ 3 %))
    (def anon-fn (fn [a b] (+ a b)))
    

    View Slide

19. Clojure
    (defn greeter [greeting]
    (fn [name]
    (println greeting “, “ name)))
    (let [german-greeting (greeter “Hallo”)]
    (german-greeting “Rich”))
    

    View Slide

20. Clojure
    (let [m {:foo 0 :bar 1}
    a [3 2 1 0]]
    (conj a 4) ;; => (4 3 2 1 0)
    (assoc m :baz 2)) ;; => {:foo 0 :bar 1 :baz
    2}
    

    View Slide

21. Clojure
    (ns my-sql-query-ns
    (:require [clojure.data.jdbc :as jdbc]
    [clojure.data.jdbc.sql :as sql]))
    (jdbc/query mysql-db
    (sql/select * :users))
    ;;=> ({:login “root” :uid 0} {:login :daemon :uid 1})
    

    View Slide

22. Clojure
    (def lotsa-ones (repeat 1)) ;; => lazy-seq
    (take 5 lotsa-ones) ;; => (1 1 1 1 1)
    (take 2
    (interleave lotsa-ones (repeat 2))) ;; => (1 2)
    

    View Slide

23. Clojure
    (def requests (atom 0))
    (swap! requests inc)
    @request ;; => 1
    (def user1 (ref 10))
    (def user2 (ref 50))
    (dosync
    (alter user1 + 10)
    (alter user2 - 10))
    @user1 ;; => 20
    @user2 ;; => 40
    

    View Slide

24. Clojure
    (def users
    {:alice {:balance 10}
    :bob {:balance 5}})
    (defn transaction [from to amount]
    (-> users
    (update-in [from :balance] - amount)
    (update-in [to :balance] + amount)))
    

    View Slide

25. Index
    (def index
    (default {:ttl 60 :state “ok”}
    (update-index (index))))
    

    View Slide

26. Streams
    (let [index (default {:ttl 60 :state “ok”}
    (update-index (index)))]
    (streams
    index))
    

    View Slide

27. Streams
    (let [index (default {:ttl 60 :state “ok”}
    (update-index (index)))
    email (mailer {:from “riemann@production”})]
    (streams
    index
    (where (state “critical”)
    email)
    ))
    

    View Slide

28. Demo!
    

    View Slide