Merge pull request #48 from DataCrunch-io/feature/instance-create-retry-params

instances.create: expose retry settings, add exponential backoff

Author: shamrin

CHANGELOG.md

@@ -7,6 +7,14 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+### Added
+
+- Added `max_wait_time`, `initial_interval`, `max_interval`, `backoff_coefficient` keyword arguments to `instances.create()`
+
+### Changed
+
+- Cap `instances.create()` retry interval to 5 seconds; add exponential backoff; increase default `max_wait_time` from 60 to 180 seconds
+
 ## [1.14.0] - 2025-08-15
 
 ### Added

datacrunch/instances/instances.py

@@ -1,4 +1,5 @@
 import time
+import itertools
 from typing import List, Union, Optional, Dict, Literal
 from dataclasses import dataclass
 from dataclasses_json import dataclass_json
@@ -123,7 +124,12 @@ def create(self,
                is_spot: bool = False,
                contract: Optional[Contract] = None,
                pricing: Optional[Pricing] = None,
-               coupon: Optional[str] = None) -> Instance:
+               coupon: Optional[str] = None,
+               *,
+               max_wait_time: float = 180,
+               initial_interval: float = 0.5,
+               max_interval: float = 5,
+               backoff_coefficient: float = 2.0) -> Instance:
         """Creates and deploys a new cloud instance.
 
         Args:
@@ -141,6 +147,10 @@ def create(self,
             contract: Optional contract type for the instance.
             pricing: Optional pricing model for the instance.
             coupon: Optional coupon code for discounts.
+            max_wait_time: Maximum total wait for the instance to start provisioning, in seconds (default: 180)
+            initial_interval: Initial interval, in seconds (default: 0.5)
+            max_interval: The longest single delay allowed between retries, in seconds (default: 5)
+            backoff_coefficient: Coefficient to calculate the next retry interval (default 2.0)
 
         Returns:
             The newly created instance object.
@@ -169,20 +179,19 @@ def create(self,
         id = self._http_client.post(INSTANCES_ENDPOINT, json=payload).text
 
         # Wait for instance to enter provisioning state with timeout
-        MAX_WAIT_TIME = 60  # Maximum wait time in seconds
-        POLL_INTERVAL = 0.5  # Time between status checks
-
-        start_time = time.time()
-        while True:
+        deadline = time.monotonic() + max_wait_time
+        for i in itertools.count():
             instance = self.get_by_id(id)
             if instance.status != InstanceStatus.ORDERED:
                 return instance
 
-            if time.time() - start_time > MAX_WAIT_TIME:
+            now = time.monotonic()
+            if now >= deadline:
                 raise TimeoutError(
-                    f"Instance {id} did not enter provisioning state within {MAX_WAIT_TIME} seconds")
+                    f"Instance {id} did not enter provisioning state within {max_wait_time:.1f} seconds")
 
-            time.sleep(POLL_INTERVAL)
+            interval = min(initial_interval * backoff_coefficient ** i, max_interval, deadline - now)
+            time.sleep(interval)
 
     def action(self, id_list: Union[List[str], str], action: str, volume_ids: Optional[List[str]] = None) -> None:
         """Performs an action on one or more instances.