Add retry logic for server creation with configurable attempts and delay

Introduce create_retries and create_retry_delay inputs to control the number of retry attempts and delay between attempts when creating a Hetzner Cloud server. Update action.sh to implement the retry loop and improve error handling for transient failures. Document new options in README.md and action.yml to enhance reliability.

Author: xgboosted

README.md

@@ -160,6 +160,12 @@ jobs:
 | `server_type`       |   | Name of the Server type this Server should be created with. | `cx22` (Intel x86, 2 vCPU, 4GB RAM, 40GB SSD) |
 | `server_wait`       |   | Wait up to `server_wait` retries (10 sec each) for the Hetzner Cloud Server to start. | `30` (5 min) |
 | `ssh_key`           |   | SSH key ID (integer) which should be injected into the Server at creation time. | `null` |
+| `create_retries`    |   | Number of retry attempts for runner creation if it fails. | `1` |
+| `create_retry_delay`|   | Delay (in seconds) between retry attempts for runner creation. | `10` |
+
+### Retry Logic
+
+If runner creation fails due to a transient error, the action will retry up to `create_retries` times, waiting `create_retry_delay` seconds between attempts. All attempts and errors are logged.
 
 ## Outputs
 

action.sh

@@ -202,7 +202,17 @@ fi
 MY_RUNNER_WAIT=${INPUT_RUNNER_WAIT:-"60"}
 # Check if MY_RUNNER_WAIT is an integer
 if [[ ! "$MY_RUNNER_WAIT" =~ ^[0-9]+$ ]]; then
-	exit_with_failure "The maximum wait time (reties) for GitHub Action Runner registration must be an integer!"
+	exit_with_failure "The maximum wait time (retries) for GitHub Action Runner registration must be an integer!"
+fi
+
+# Set create retry logic inputs
+MY_CREATE_RETRIES=${INPUT_CREATE_RETRIES:-1}
+MY_CREATE_RETRY_DELAY=${INPUT_CREATE_RETRY_DELAY:-10}
+if [[ ! "$MY_CREATE_RETRIES" =~ ^[0-9]+$ ]]; then
+	exit_with_failure "The create_retries value must be an integer!"
+fi
+if [[ ! "$MY_CREATE_RETRY_DELAY" =~ ^[0-9]+$ ]]; then
+	exit_with_failure "The create_retry_delay value must be an integer!"
 fi
 
 # Set Hetzner Cloud Server ID
@@ -363,18 +373,31 @@ fi
 
 # Send a POST request to the Hetzner Cloud API to create a server.
 # https://docs.hetzner.cloud/#servers-create-a-server
-echo "Create server..."
-if ! curl \
+MY_CREATE_ATTEMPT=1
+while [[ $MY_CREATE_ATTEMPT -le $MY_CREATE_RETRIES ]]; do
+	echo "Create server (attempt $MY_CREATE_ATTEMPT of $MY_CREATE_RETRIES)..."
+	if curl \
 	-X POST \
 	--fail-with-body \
 	-o "servers.json" \
 	-H "Content-Type: application/json" \
 	-H "Authorization: Bearer ${MY_HETZNER_TOKEN}" \
 	-d @create-server.json \
 	"https://api.hetzner.cloud/v1/servers"; then
-	cat "servers.json"
-	exit_with_failure "Failed to create Server in Hetzner Cloud!"
-fi
+		echo "Server created successfully."
+		break
+	else
+		echo "Server creation failed (attempt $MY_CREATE_ATTEMPT)."
+		if [[ $MY_CREATE_ATTEMPT -lt $MY_CREATE_RETRIES ]]; then
+			echo "Retrying in $MY_CREATE_RETRY_DELAY seconds..."
+			sleep "$MY_CREATE_RETRY_DELAY"
+		else
+			cat "servers.json"
+			exit_with_failure "Failed to create Server in Hetzner Cloud after $MY_CREATE_RETRIES attempts!"
+		fi
+	fi
+	MY_CREATE_ATTEMPT=$((MY_CREATE_ATTEMPT + 1))
+done
 
 # Get the Hetzner Server ID from the JSON response (assuming valid JSON)
 MY_HETZNER_SERVER_ID=$(jq -er '.server.id' < "servers.json")

action.yml

@@ -102,6 +102,16 @@ inputs:
       Specifies bash commands to run before the GitHub Actions Runner starts.
       It's useful for installing dependencies with apt-get, dnf, zypper etc.
     required: false
+  create_retries:
+    description: >-
+      Number of retry attempts for runner creation if it fails.
+    required: false
+    default: '1'
+  create_retry_delay:
+    description: >-
+      Delay (in seconds) between retry attempts for runner creation.
+    required: false
+    default: '10'
 
 outputs:
   label:
@@ -142,3 +152,5 @@ runs:
         INPUT_SERVER_TYPE: ${{ inputs.server_type }}
         INPUT_SERVER_WAIT: ${{ inputs.server_wait }}
         INPUT_SSH_KEY: ${{ inputs.ssh_key }}
+        INPUT_CREATE_RETRIES: ${{ inputs.create_retries }}
+        INPUT_CREATE_RETRY_DELAY: ${{ inputs.create_retry_delay }}