address cron job confusion and update dependencies

CHANGELOG.md

@@ -1,30 +1,36 @@
 # pod-reaper: kills pods dead
 
+### 2.6.0
+- adjusted cron `SCHEDULE` for optional seconds, non optional day of week
+
+### 2.5.0
+- added multiple logging formats
+
+### 2.4.1
+- added configurable `LOG_LEVEL`
+
+### 2.4.0
+- added `UNREADY` rule to kill pods based on duration of time not passing readiness checks
+
 ### 2.3.0
 - added `POD_STATUSES` rule (can now filter/kill `Evicted` pods)
 
 ### 2.2.0
-
 - added configurable `GRACE_PERIOD` to control soft vs hard pod kills
 
 ### 2.1.0
-
 - Added logging via [logrus](https://github.com/sirupsen/logrus)
 
 ## 2.0.0
-
 - removed `POLL_INTERVAL` environment variable in favor of cron schedule
 - added `SCHEDULE` environment variable to control when pods are inspected for reaping
   - makes use of https://godoc.org/github.com/robfig/cron
 - refactored packages for clarity
 - testing refactor for clarity
-
 ### 1.1.0
-
 - added ability to only reap pods with specified labels
 
 ## 1.0.0
-
 - redesign of the reaper to be built on modular rules
   - rules must implement two methods `load()` and `shouldReap(pod)`
   - rules determine whether or not they get loaded at runtime via environment variables

Dockerfile

@@ -1,14 +1,12 @@
 # Build
-FROM golang:1.9 AS build
+FROM golang:1.14 AS build
 WORKDIR /go/src/github.com/target/pod-reaper
 ENV CGO_ENABLED=0 GOOS=linux
-RUN go get github.com/Masterminds/glide
-COPY glide.* ./
-RUN glide install --strip-vendor
-COPY reaper/*.go ./reaper/
-COPY rules/*.go ./rules/
-RUN go test $(glide nv)
-RUN go build -o pod-reaper -a -installsuffix go ./reaper
+COPY ./ ./
+RUN go get github.com/Masterminds/glide \
+ && glide install --strip-vendor \
+ && go test $(glide nv) \
+ && go build -o pod-reaper -a -installsuffix go ./reaper
 
 # Application
 FROM scratch

README.md

@@ -51,7 +51,7 @@ Controls the grace period between a soft pod termination and a hard termination.
 
 Default value: "@every 1m"
 
-Controls how frequently pod-reaper queries kubernetes for pods. The format follows the upstream cron library https://godoc.org/github.com/robfig/cron. For most use cases, the interval format `@every 1h2m3s` is sufficient. But more complex use cases can make use of the `* * * * *` notation.
+Controls how frequently pod-reaper queries kubernetes for pods. The format follows the upstream cron library https://godoc.org/github.com/robfig/cron. For most use cases, the interval format `@every 1h2m3s` is sufficient. But more complex use cases can make use of the `* * * * *` notation. The cron parser used can optionally support seconds if a sixth parameter is add. `12 * * * * *` for example will run on the 12th second of every minute.
 
 ### `RUN_DURATION`
 

examples/complex-deployment.yml

@@ -99,6 +99,9 @@ spec:
           # randomly flag 30% of pods whenever they are checked
           - name: CHAOS_CHANCE
             value: ".3"
+          # increase logging
+          - name: LOG_LEVEL
+            value: debug
 
       - name: chaos # reaper 2
         image: target/pod-reaper
@@ -110,12 +113,15 @@ spec:
             cpu: 20m
             memory: 20Mi
         env:
-          # cron job based schedule
+          # cron job based schedule (with seconds)
           - name: SCHEDULE
-            value: "0/5 * * * *"
+            value: "15 * * * * *"
           # randomly flag 5% of pods whenever they are checked
           - name: CHAOS_CHANCE
             value: ".05"
+          # increase logging
+          - name: LOG_LEVEL
+            value: debug
 
       - name: error # reaper 3
         image: target/pod-reaper
@@ -130,7 +136,10 @@ spec:
         env:
           # check pods every 3 min
           - name: SCHEDULE
-            value: "@every 3m"
+            value: "0/3 * * * *"
           # check if container status is in the set  { Error, ErrImagePull, ImagePullBackOff }
           - name: CONTAINER_STATUSES
             value: "Error,ErrImagePull,ImagePullBackOff"
+          # increase logging
+          - name: LOG_LEVEL
+            value: debug