Add validation to cw url

Author: zhaoqizqwang

src/hyperpod_cli/service/get_logs.py

@@ -23,9 +23,11 @@
 from hyperpod_cli.utils import (
     get_eks_cluster_name,
     get_hyperpod_cluster_region,
+    validate_region_and_cluster_name,
 )
 from kubernetes.client.rest import ApiException
 from kubernetes.client import V1ResourceAttributes
+import re
 
 AMAZON_ClOUDWATCH_OBSERVABILITY = "amazon-cloudwatch-observability"
 
@@ -103,18 +105,29 @@ def generate_cloudwatch_link(
                 container_id = None         
 
             # Cloudwatch container insight log groups should have the same pod log as API response
-            if node_name and pod_name and namespace and container_name and container_id:
-                region = get_hyperpod_cluster_region()
+            region = get_hyperpod_cluster_region()
+            cloudwatch_url = self.get_log_url(eks_cluster_name, region, node_name, pod_name, namespace, container_name, container_id)
 
-                cloudwatch_url = self.get_log_url(eks_cluster_name, region, node_name, pod_name, namespace, container_name, container_id)
-                cloudwatch_link = f'The pod cloudwatch log stream link is {cloudwatch_url}'
-            else:
-                cloudwatch_link = 'Failed to load container insights CloudWatch Link!'
+            if not validate_region_and_cluster_name(region, eks_cluster_name):
+                return 'Failed to validate Eks cluster name and region'
+
+            if not self._validate_log_url(cloudwatch_url):
+                return 'Failed to validate cloudwatch log url. Please verify node name, pod container name and container id are valid'
+            
+            cloudwatch_link = f'The pod cloudwatch log stream link is {cloudwatch_url}'
         else:
             cloudwatch_link = None
 
         return cloudwatch_link
 
+    def _validate_log_url(self, log_url):
+        pattern = "https:\/\/([a-z0-9-]+).console.aws.amazon.com\/cloudwatch\/home\?region=([a-z0-9-]+)#logsV2:log-groups\/log-group\/\$252Faws\$252Fcontainerinsights\$252F([a-zA-Z0-9-]+)\$252Fapplication\/log-events\/([a-z0-9-]+)-application.var.log.containers.([a-z0-9-]+)_([a-z0-9-]+)_([a-z0-9-]+)-([a-z0-9-]+).log"
+        match = re.match(pattern, log_url)
+        if match:
+            return True
+        else:
+            return False
+
     def get_log_url(self, eks_cluster_name, region, node_name, pod_name, namespace, container_name, container_id):
         console_prefix = f'https://{region}.console.aws.amazon.com/cloudwatch/home?region={region}#'
         log_group_prefix = f'logsV2:log-groups/log-group/$252Faws$252Fcontainerinsights$252F{eks_cluster_name}$252Fapplication/log-events/'

src/hyperpod_cli/utils.py

@@ -121,7 +121,7 @@ def _validate_link(console_url):
         return False
 
 
-def _validate_placeholders(region, cluster_name):
+def validate_region_and_cluster_name(region, cluster_name):
     output = False
     region_char_list = region.split("-")
 
@@ -135,6 +135,9 @@ def _validate_placeholders(region, cluster_name):
     region_prefix_length = len(region_char_list[0])
     region_length = len(region_char_list[1])
     region_suffix_length = len(region_char_list[2])
+
+    cluster_name_match = re.match("[a-zA-Z0-9-]+", cluster_name)
+
     if (
         region_prefix_match
         and region_match
@@ -143,6 +146,7 @@ def _validate_placeholders(region, cluster_name):
         and region_suffix_length == 1
         and region_length >= 4
         and region_length < 10
+        and cluster_name_match
         and len(cluster_name) >= 1
         and len(cluster_name) <= 63
     ):
@@ -165,7 +169,7 @@ def get_cluster_console_url():
             f"https://{region}.console.aws.amazon.com/sagemaker/"
             f"home?region={region}#/cluster-management/{cluster_name}"
         )
-        if _validate_link(console_url) and _validate_placeholders(region, cluster_name):
+        if _validate_link(console_url) and validate_region_and_cluster_name(region, cluster_name):
             return console_url
     return None