Add support for fleets created through the CreateFleet API (#410)

* Added support for the CreateFleet API * Fixed EC2CreateFleet usage of launch template overrides for retreiving instance type and weight capacity * Rename all occurances of EC2Fleet to Fleet, and rename CreateFleet to EC2Fleet * Renaming instances of 'EC2Fleet' to 'Fleet' * Add permision checks and add missing permissions for EC2Fleets, fleets created by the CreateFleet API --------- Co-authored-by: Prathibha Datta Kumar <[email protected]>
jenkinsci · Nov 2, 2023 · 9aaaac1 · 9aaaac1
1 parent d79c241
commit 9aaaac1
Show file tree

Hide file tree

Showing 72 changed files with 1,324 additions and 743 deletions.
diff --git a/README.md b/README.md
@@ -208,7 +208,7 @@ import hudson.plugins.sshslaves.verifiers.NonVerifyingKeyVerificationStrategy
 import com.cloudbees.plugins.credentials.*
 import com.cloudbees.plugins.credentials.domains.Domain
 import hudson.model.*
-import com.amazon.jenkins.ec2fleet.EC2FleetCloud
+import com.amazon.jenkins.ec2fleet.FleetCloud
 import jenkins.model.Jenkins
 
 // just modify this config other code just logic
@@ -247,7 +247,7 @@ BasicSSHUserPrivateKey instanceCredentials = new BasicSSHUserPrivateKey(
 )
 // find detailed information about parameters on plugin config page or
 // https://github.com/jenkinsci/ec2-fleet-plugin/blob/master/src/main/java/com/amazon/jenkins/ec2fleet/EC2FleetCloud.java
-EC2FleetCloud ec2FleetCloud = new EC2FleetCloud(
+FleetCloud fleetCloud = new FleetCloud(
   "", // fleetCloudName
   null,
   awsCredentials.id,

diff --git a/docs/CONFIGURATION-AS-CODE.md b/docs/CONFIGURATION-AS-CODE.md
@@ -4,9 +4,9 @@
 
 ## Properties
 
-### EC2FleetCloud
+### FleetCloud
 
-[Definition](https://github.com/jenkinsci/ec2-fleet-plugin/blob/master/src/main/java/com/amazon/jenkins/ec2fleet/EC2FleetCloud.java#L156-L179)
+[Definition](https://github.com/jenkinsci/ec2-fleet-plugin/blob/master/src/main/java/com/amazon/jenkins/ec2fleet/FleetCloud.java#L156-L179)
 
 | Property                   | Type    | Required                                 | Description                                                                                                                       |
 |----------------------------|---------|------------------------------------------|-----------------------------------------------------------------------------------------------------------------------------------|
@@ -35,20 +35,20 @@
 | cloudStatusIntervalSec     | int     | no, default ```10```                     ||
 | noDelayProvision           | boolean | no, default ```false```                  ||
 
-## EC2FleetLabelCloud
+## FleetLabelCloud
 
 More about this type [here](LABEL-BASED-CONFIGURATION.md)
 
-[Definition](https://github.com/jenkinsci/ec2-fleet-plugin/blob/master/src/main/java/com/amazon/jenkins/ec2fleet/EC2FleetLabelCloud.java#L123-L145)
+[Definition](https://github.com/jenkinsci/ec2-fleet-plugin/blob/master/src/main/java/com/amazon/jenkins/ec2fleet/FleetLabelCloud.java#L123-L145)
 
 ## Examples
 
-### EC2FleetCloud (min set of properties)
+### FleetCloud (min set of properties)
 
 ```yaml
 jenkins:
   clouds:
-    - ec2Fleet:
+    - eC2Fleet:
         name: ec2-fleet
         computerConnector:
             sshConnector:
@@ -61,12 +61,12 @@ jenkins:
         maxSize: 10
 ```
 
-### EC2FleetCloud (All properties)
+### FleetCloud (All properties)
 
 ```yaml
 jenkins:
   clouds:
-    - ec2Fleet:
+    - eC2Fleet:
         name: ec2-fleet
         awsCredentialsId: xx
         computerConnector:

diff --git a/pom.xml b/pom.xml
@@ -57,6 +57,11 @@
       <name>Steve Nay</name>
       <email>[email protected]</email>
     </developer>
+    <developer>
+      <id>GavinBurris42</id>
+      <name>Gavin Burris</name>
+      <email>[email protected]</email>
+    </developer>
   </developers>
 
   <scm>

diff --git a/...nkins/ec2fleet/AbstractEC2FleetCloud.java → .../jenkins/ec2fleet/AbstractFleetCloud.java b/...nkins/ec2fleet/AbstractEC2FleetCloud.java → .../jenkins/ec2fleet/AbstractFleetCloud.java
@@ -2,9 +2,9 @@
 
 import hudson.slaves.Cloud;
 
-public abstract class AbstractEC2FleetCloud extends Cloud {
+public abstract class AbstractFleetCloud extends Cloud {
 
-    protected AbstractEC2FleetCloud(String name) {
+    protected AbstractFleetCloud(String name) {
         super(name);
     }
 

diff --git a/src/main/java/com/amazon/jenkins/ec2fleet/CloudNames.java b/src/main/java/com/amazon/jenkins/ec2fleet/CloudNames.java
@@ -1,6 +1,5 @@
 package com.amazon.jenkins.ec2fleet;
 
-import hudson.slaves.Cloud;
 import jenkins.model.Jenkins;
 import org.apache.commons.lang.RandomStringUtils;
 
@@ -26,7 +25,7 @@ public static String generateUnique(final String proposedName) {
             ? Jenkins.get().clouds.stream().map(c -> c.name).collect(Collectors.toSet())
             : Collections.emptySet();
 
-    if (proposedName.equals(EC2FleetCloud.BASE_DEFAULT_FLEET_CLOUD_ID) || proposedName.equals(EC2FleetLabelCloud.BASE_DEFAULT_FLEET_CLOUD_ID) || usedNames.contains(proposedName)) {
+    if (proposedName.equals(FleetCloud.BASE_DEFAULT_FLEET_CLOUD_ID) || proposedName.equals(FleetLabelCloud.BASE_DEFAULT_FLEET_CLOUD_ID) || usedNames.contains(proposedName)) {
       return proposedName + "-" + generateSuffix();
     }
 

diff --git a/src/main/java/com/amazon/jenkins/ec2fleet/CloudNanny.java b/src/main/java/com/amazon/jenkins/ec2fleet/CloudNanny.java
@@ -14,7 +14,7 @@
 import java.util.logging.Logger;
 
 /**
- * {@link CloudNanny} is responsible for periodically running update (i.e. sync-state-with-AWS) cycles for {@link EC2FleetCloud}s.
+ * {@link CloudNanny} is responsible for periodically running update (i.e. sync-state-with-AWS) cycles for {@link FleetCloud}s.
  */
 @Extension
 @SuppressWarnings("unused")
@@ -23,7 +23,7 @@ public class CloudNanny extends PeriodicWork {
     private static final Logger LOGGER = Logger.getLogger(CloudNanny.class.getName());
 
     // the map should not hold onto fleet instances to allow deletion of fleets.
-    private final Map<EC2FleetCloud, AtomicInteger> recurrenceCounters = Collections.synchronizedMap(new WeakHashMap<>());
+    private final Map<FleetCloud, AtomicInteger> recurrenceCounters = Collections.synchronizedMap(new WeakHashMap<>());
 
     @Override
     public long getRecurrencePeriod() {
@@ -39,8 +39,8 @@ public long getRecurrencePeriod() {
     @Override
     protected void doRun() {
         for (final Cloud cloud : getClouds()) {
-            if (!(cloud instanceof EC2FleetCloud)) continue;
-            final EC2FleetCloud fleetCloud = (EC2FleetCloud) cloud;
+            if (!(cloud instanceof FleetCloud)) continue;
+            final FleetCloud fleetCloud = (FleetCloud) cloud;
 
             final AtomicInteger recurrenceCounter = getRecurrenceCounter(fleetCloud);
 
@@ -70,7 +70,7 @@ private static List<Cloud> getClouds() {
         return Jenkins.get().clouds;
     }
 
-    private AtomicInteger getRecurrenceCounter(EC2FleetCloud fleetCloud) {
+    private AtomicInteger getRecurrenceCounter(FleetCloud fleetCloud) {
         AtomicInteger counter = new AtomicInteger(fleetCloud.getCloudStatusIntervalSec());
         // If a counter already exists, return the value, otherwise set the new counter value and return it.
         AtomicInteger existing = recurrenceCounters.putIfAbsent(fleetCloud, counter);

diff --git a/src/main/java/com/amazon/jenkins/ec2fleet/EC2RetentionStrategy.java b/src/main/java/com/amazon/jenkins/ec2fleet/EC2RetentionStrategy.java
@@ -13,24 +13,24 @@
 import java.util.logging.Logger;
 
 /**
- * The {@link EC2RetentionStrategy} controls when to take {@link EC2FleetNodeComputer} offline, bring it back online, or even to destroy it.
+ * The {@link EC2RetentionStrategy} controls when to take {@link FleetNodeComputer} offline, bring it back online, or even to destroy it.
  */
-public class EC2RetentionStrategy extends RetentionStrategy<EC2FleetNodeComputer> implements ExecutorListener {
+public class EC2RetentionStrategy extends RetentionStrategy<FleetNodeComputer> implements ExecutorListener {
     private static final Logger LOGGER = Logger.getLogger(EC2RetentionStrategy.class.getName());
     private static final int RE_CHECK_IN_A_MINUTE = 1;
 
     /**
      * Will be called under {@link hudson.model.Queue#withLock(Runnable)}
      *
-     * @param fc EC2FleetNodeComputer
+     * @param fc FleetNodeComputer
      * @return delay in min before next run
      */
     @SuppressFBWarnings(
             value = "BC_UNCONFIRMED_CAST",
-            justification = "to ignore EC2FleetNodeComputer cast")
+            justification = "to ignore FleetNodeComputer cast")
     @Override
-    public long check(final EC2FleetNodeComputer fc) {
-        final AbstractEC2FleetCloud cloud = fc.getCloud();
+    public long check(final FleetNodeComputer fc) {
+        final AbstractFleetCloud cloud = fc.getCloud();
 
         LOGGER.fine(String.format("Checking if node '%s' is idle ", fc.getName()));
 
@@ -41,7 +41,7 @@ public long check(final EC2FleetNodeComputer fc) {
             return RE_CHECK_IN_A_MINUTE;
         }
 
-        // Ensure that the EC2FleetCloud cannot be mutated from under us while
+        // Ensure that the FleetCloud cannot be mutated from under us while
         // we're doing this check
         // Ensure nobody provisions onto this node until we've done
         // checking
@@ -58,14 +58,14 @@ public long check(final EC2FleetNodeComputer fc) {
                 EC2AgentTerminationReason reason;
                 // Determine the reason for termination from specific to generic use cases.
                 // Reasoning for checking all cases of termination initiated by the plugin:
-                //  A user-initiated change to cloud configuration creates a new EC2FleetCloud object, erasing class fields containing data like instance IDs to terminate.
+                //  A user-initiated change to cloud configuration creates a new FleetCloud object, erasing class fields containing data like instance IDs to terminate.
                 //  Hence, determine the reasons for termination here using persisted fields for accurate handling of termination.
                 if (fc.isMarkedForDeletion()) {
                     reason = EC2AgentTerminationReason.AGENT_DELETED;
                 } else if (cloud.hasExcessCapacity()) {
                     reason = EC2AgentTerminationReason.EXCESS_CAPACITY;
-                } else if (cloud instanceof EC2FleetCloud && !((EC2FleetCloud) cloud).hasUnlimitedUsesForNodes()
-                        && ((EC2FleetNode)node).getUsesRemaining() <= 0) {
+                } else if (cloud instanceof FleetCloud && !((FleetCloud) cloud).hasUnlimitedUsesForNodes()
+                        && ((FleetNode)node).getUsesRemaining() <= 0) {
                     reason = EC2AgentTerminationReason.MAX_TOTAL_USES_EXHAUSTED;
                 } else if (isIdleForTooLong(cloud, fc)) {
                     reason = EC2AgentTerminationReason.IDLE_FOR_TOO_LONG;
@@ -97,12 +97,12 @@ public long check(final EC2FleetNodeComputer fc) {
     }
 
     @Override
-    public void start(EC2FleetNodeComputer c) {
+    public void start(FleetNodeComputer c) {
         LOGGER.log(Level.INFO, "Connecting to instance: " + c.getDisplayName());
         c.connect(false);
     }
 
-    private boolean isIdleForTooLong(final AbstractEC2FleetCloud cloud, final Computer computer) {
+    private boolean isIdleForTooLong(final AbstractFleetCloud cloud, final Computer computer) {
         final int idleMinutes = cloud.getIdleMinutes();
         if (idleMinutes <= 0) return false;
 
@@ -122,25 +122,25 @@ private boolean isIdleForTooLong(final AbstractEC2FleetCloud cloud, final Comput
 
     @Override
     public void taskAccepted(Executor executor, Queue.Task task) {
-        final EC2FleetNodeComputer computer = (EC2FleetNodeComputer) executor.getOwner();
+        final FleetNodeComputer computer = (FleetNodeComputer) executor.getOwner();
         if (computer != null) {
-            final EC2FleetNode ec2FleetNode = computer.getNode();
-            if (ec2FleetNode != null) {
-                final int maxTotalUses = ec2FleetNode.getMaxTotalUses();
+            final FleetNode fleetNode = computer.getNode();
+            if (fleetNode != null) {
+                final int maxTotalUses = fleetNode.getMaxTotalUses();
                 if (maxTotalUses <= -1) { // unlimited uses
                     LOGGER.fine("maxTotalUses set to unlimited (" + maxTotalUses + ") for agent " + computer.getName());
                 } else { // limited uses
-                    if (ec2FleetNode.getUsesRemaining() > 1) {
-                        ec2FleetNode.decrementUsesRemaining();
-                        LOGGER.info("Agent " + computer.getName() + " has " + ec2FleetNode.getUsesRemaining() + " builds left");
-                    } else if (ec2FleetNode.getUsesRemaining() == 1) { // current task should be the last task for this agent
+                    if (fleetNode.getUsesRemaining() > 1) {
+                        fleetNode.decrementUsesRemaining();
+                        LOGGER.info("Agent " + computer.getName() + " has " + fleetNode.getUsesRemaining() + " builds left");
+                    } else if (fleetNode.getUsesRemaining() == 1) { // current task should be the last task for this agent
                         LOGGER.info(String.format("maxTotalUses drained - suspending agent %s after current build", computer.getName()));
                         computer.setAcceptingTasks(false);
-                        ec2FleetNode.decrementUsesRemaining();
+                        fleetNode.decrementUsesRemaining();
                     } else {
                         // don't decrement when usesRemaining=0, as -1 has a special meaning.
                         LOGGER.warning(String.format("Agent %s accepted a task after being suspended!!! MaxTotalUses: %d, uses remaining: %d",
-                                computer.getName(), ec2FleetNode.getMaxTotalUses(), ec2FleetNode.getUsesRemaining()));
+                                computer.getName(), fleetNode.getMaxTotalUses(), fleetNode.getUsesRemaining()));
                     }
                 }
             }
@@ -170,15 +170,15 @@ private void postJobAction(final Executor executor, final Throwable throwable) {
                     TimeUnit.MILLISECONDS.toSeconds(executor.getElapsedTime())));
         }
 
-        final EC2FleetNodeComputer computer = (EC2FleetNodeComputer) executor.getOwner();
+        final FleetNodeComputer computer = (FleetNodeComputer) executor.getOwner();
         if (computer != null) {
-            final EC2FleetNode ec2FleetNode = computer.getNode();
-            if (ec2FleetNode != null) {
-                final AbstractEC2FleetCloud cloud = ec2FleetNode.getCloud();
+            final FleetNode fleetNode = computer.getNode();
+            if (fleetNode != null) {
+                final AbstractFleetCloud cloud = fleetNode.getCloud();
                 if (computer.countBusy() <= 1 && !computer.isAcceptingTasks()) {
-                    LOGGER.info("Calling scheduleToTerminate for node " + ec2FleetNode.getNodeName() + " due to exhausted maxTotalUses.");
+                    LOGGER.info("Calling scheduleToTerminate for node " + fleetNode.getNodeName() + " due to exhausted maxTotalUses.");
                     // Schedule instance for termination even if it breaches minSize and minSpareSize constraints
-                    cloud.scheduleToTerminate(ec2FleetNode.getNodeName(), true, EC2AgentTerminationReason.MAX_TOTAL_USES_EXHAUSTED);
+                    cloud.scheduleToTerminate(fleetNode.getNodeName(), true, EC2AgentTerminationReason.MAX_TOTAL_USES_EXHAUSTED);
                 }
             }
         }

diff --git a/...EC2FleetAutoResubmitComputerLauncher.java → ...et/FleetAutoResubmitComputerLauncher.java b/...EC2FleetAutoResubmitComputerLauncher.java → ...et/FleetAutoResubmitComputerLauncher.java
@@ -22,31 +22,31 @@
 import java.util.logging.Logger;
 
 /**
- * The {@link EC2FleetAutoResubmitComputerLauncher} is responsible for controlling:
- *  * how {@link EC2FleetNodeComputer}s are launched
- *  * how {@link EC2FleetNodeComputer}s connect to agents {@link EC2FleetNode}
+ * The {@link FleetAutoResubmitComputerLauncher} is responsible for controlling:
+ *  * how {@link FleetNodeComputer}s are launched
+ *  * how {@link FleetNodeComputer}s connect to agents {@link FleetNode}
  *
  * This is wrapper for {@link ComputerLauncher} to get notification when agent was disconnected
  * and automatically resubmit {@link hudson.model.Queue.Task} if reason is unexpected termination
  * which usually means EC2 instance was interrupted.
  * <p>
  * This is optional feature, it's enabled by default, but could be disabled by
- * {@link EC2FleetCloud#isDisableTaskResubmit()}
+ * {@link FleetCloud#isDisableTaskResubmit()}
  */
 @SuppressWarnings("WeakerAccess")
 @ThreadSafe
-public class EC2FleetAutoResubmitComputerLauncher extends DelegatingComputerLauncher {
+public class FleetAutoResubmitComputerLauncher extends DelegatingComputerLauncher {
 
     private static final Level LOG_LEVEL = Level.INFO;
-    private static final Logger LOGGER = Logger.getLogger(EC2FleetAutoResubmitComputerLauncher.class.getName());
+    private static final Logger LOGGER = Logger.getLogger(FleetAutoResubmitComputerLauncher.class.getName());
 
     /**
      * Delay which will be applied when job {@link Queue#scheduleInternal(Queue.Task, int, List)}
      * rescheduled after offline
      */
     private static final int RESCHEDULE_QUIET_PERIOD_SEC = 10;
 
-    public EC2FleetAutoResubmitComputerLauncher(final ComputerLauncher launcher) {
+    public FleetAutoResubmitComputerLauncher(final ComputerLauncher launcher) {
         super(launcher);
     }
 
@@ -72,14 +72,14 @@ public EC2FleetAutoResubmitComputerLauncher(final ComputerLauncher launcher) {
      */
     @SuppressFBWarnings(
             value = "BC_UNCONFIRMED_CAST",
-            justification = "to ignore EC2FleetNodeComputer cast")
+            justification = "to ignore FleetNodeComputer cast")
     @Override
     public void afterDisconnect(final SlaveComputer computer, final TaskListener listener) {
         // according to jenkins docs could be null in edge cases, check ComputerLauncher.afterDisconnect
         if (computer == null) return;
 
         // in some multi-thread edge cases cloud could be null for some time, just be ok with that
-        final AbstractEC2FleetCloud cloud = ((EC2FleetNodeComputer) computer).getCloud();
+        final AbstractFleetCloud cloud = ((FleetNodeComputer) computer).getCloud();
         if (cloud == null) {
             LOGGER.warning("Cloud is null for computer " + computer.getDisplayName()
                     + ". This should be autofixed in a few minutes, if not please create an issue for the plugin");