Merge pull request 'rate-limit' (#3) from rate-limit into main

Limiting the rates

Introducing rate limiting seemed like a very sensible step in order to mitigate against some forms of DoS attacks. As of now, we set the rates to a fixed amount with no configuration ability. This might change later on, depending on the use cases.

build.py

@@ -56,7 +56,7 @@ def test_schema(project):
         "java -jar target/uberjar/c4k-forgejo-standalone.jar "
         + "src/test/resources/forgejo-test/valid-config.yaml "
         + "src/test/resources/forgejo-test/valid-auth.yaml | "
-        + "kubeconform --kubernetes-version 1.23.0 --strict --skip Certificate -",
+        + """kubeconform --kubernetes-version 1.23.0 --strict --skip "Certificate, Middleware" -""",
         shell=True,
         check=True,
     )

src/main/cljc/dda/c4k_forgejo/core.cljc

@@ -9,6 +9,7 @@
   [dda.c4k-common.postgres :as postgres]))
 
 (def config-defaults {:issuer "staging", :deploy-federated "false"})
+(def rate-limit-defaults {:max-rate 10, :max-concurrent-requests 5})
 
 (def config? (s/keys :req-un [::forgejo/fqdn 
                               ::forgejo/mailer-from 
@@ -30,7 +31,7 @@
 
 (def vol? (s/keys :req-un [::forgejo/volume-total-storage-size]))
 
-(defn k8s-objects [config auth]
+(defn k8s-objects [config auth] ; ToDo: ADR for generate functions - vector or no vector?
   (let [storage-class (if (contains? config :postgres-data-volume-path) :manual :local-path)]
     (map yaml/to-string
          (filter #(not (nil? %))
@@ -49,8 +50,9 @@
                    (forgejo/generate-service-ssh)
                    (forgejo/generate-data-volume config)
                    (forgejo/generate-appini-env config)
-                   (forgejo/generate-secrets auth)] 
+                   (forgejo/generate-secrets auth)
-                  (forgejo/generate-ingress-and-cert config)
+                   (forgejo/generate-rate-limit-middleware rate-limit-defaults)] ; this does not have a vector as output
+                  (forgejo/generate-rate-limit-ingress-and-cert config) ; this function has a vector as output
                   (when (contains? config :restic-repository)
                     [(backup/generate-config config)
                      (backup/generate-secret auth)

src/main/cljc/dda/c4k_forgejo/forgejo.cljc

@@ -42,6 +42,8 @@
 (s/def ::mailer-pw pred/bash-env-string?)
 (s/def ::issuer pred/letsencrypt-issuer?)
 (s/def ::volume-total-storage-size (partial pred/int-gt-n? 5))
+(s/def ::max-rate int?) 
+(s/def ::max-concurrent-requests int?)
 
 (def config? (s/keys :req-un [::fqdn
                               ::mailer-from
@@ -53,6 +55,9 @@
                               ::default-app-name
                               ::service-domain-whitelist]))
 
+(def rate-limit-config? (s/keys :req-un [::max-rate
+                                         ::max-concurrent-requests]))
+
 (def auth? (s/keys :req-un [::postgres/postgres-db-user ::postgres/postgres-db-password ::mailer-user ::mailer-pw]))
 
 (def vol? (s/keys :req-un [::volume-total-storage-size]))
@@ -119,6 +124,26 @@
        :fqdns [fqdn]}
       config))))
 
+(defn-spec generate-rate-limit-ingress-and-cert pred/map-or-seq?
+  [config config?]
+  (->
+   (generate-ingress-and-cert config) ; returns a vector
+   (#(assoc-in % ; Attention: heavily relying on the output order of ing/generate-ingress-and-cert
+               [1 :metadata :annotations :traefik.ingress.kubernetes.io/router.middlewares]
+               (str
+                (-> (second %) :metadata :annotations :traefik.ingress.kubernetes.io/router.middlewares)
+                ", default-ratelimit@kubernetescrd")))))
+
+
+; using :average and :burst seems sensible, :period may be interesting for fine tuning later on
+(defn-spec generate-rate-limit-middleware pred/map-or-seq?
+  [config rate-limit-config?]
+  (let [{:keys [max-rate max-concurrent-requests]} config]
+  (->
+   (yaml/load-as-edn "forgejo/middleware-ratelimit.yaml")
+   (cm/replace-key-value :average max-rate)
+   (cm/replace-key-value :burst max-concurrent-requests))))
+
 (defn-spec generate-data-volume pred/map-or-seq?
   [config vol?]
   (let [{:keys [volume-total-storage-size]} config        

src/main/resources/forgejo/middleware-ratelimit.yaml

@@ -0,0 +1,8 @@
+apiVersion: traefik.containo.us/v1alpha1
+kind: Middleware
+metadata:
+  name: ratelimit
+spec:
+  rateLimit: # Config options for rate limiting: https://doc.traefik.io/traefik/middlewares/http/ratelimit/
+    average: AVG
+    burst: BRS

src/test/cljc/dda/c4k_forgejo/forgejo_test.cljc

@@ -130,3 +130,26 @@
           :storage-c2 "15Gi"}
          (th/map-diff (cut/generate-data-volume {:volume-total-storage-size 1})
                       (cut/generate-data-volume {:volume-total-storage-size 15})))))
+
+(deftest should-generate-middleware-ratelimit
+  (is (= {:apiVersion "traefik.containo.us/v1alpha1",
+          :kind "Middleware",
+          :metadata {:name "ratelimit"},
+          :spec {:rateLimit {:average 10, :burst 5}}}
+         (cut/generate-rate-limit-middleware {:max-rate 10, :max-concurrent-requests 5}))))
+
+(deftest should-generate-middleware-ratelimit-ingress-and-cert
+  (is (= {:traefik.ingress.kubernetes.io/router.entrypoints "web, websecure",
+          :traefik.ingress.kubernetes.io/router.middlewares
+          "default-redirect-https@kubernetescrd, default-ratelimit@kubernetescrd",
+          :metallb.universe.tf/address-pool "public"}
+         (-> (second
+              (cut/generate-rate-limit-ingress-and-cert
+               {:fqdn "test.de"
+                :mailer-from ""
+                :mailer-host "m.t.de"
+                :mailer-port "123"
+                :service-noreply-address ""
+                :average 10
+                :burst 5}))
+             :metadata :annotations))))