Error: There is no container instance that can be provided to this Durable Object, try again later

[alexander-zuev]

Issue Summary

Container instances are not available on production Cloudflare Workers account, causing all container-related operations to fail with error: "There is no container instance that can be
provided to this Durable Object, try again later"

Environment Details

• Account: Production Cloudflare Workers account (can share per request)
• Wrangler Version: 4.23.0
• Container Config: instance_type: "dev" with @cloudflare/containers SDK
• Deployment: Working locally and in tutorial examples, failing in production

Error Details

ERROR: There is no container instance that can be provided to this Durable Object, try again later

Configuration wrangler.jsonc:

  {
    "containers": [{
      "name": "code-container",
      "class_name": "CodeContainerDO",
      "image": "./Dockerfile",
      "instance_type": "dev"
    }],
    "durable_objects": {
      "bindings": [{
        "class_name": "CodeContainerDO",
        "name": "CODE_PORTABLE"
      }]
    }
  }

Steps to Reproduce

1. Deploy worker with container configuration
2. Call any endpoint that uses containerDO.start()
3. Observe error in logs

Expected Behavior

Container should start successfully as it does in tutorial examples and local development.

Current Status

• ✅ Image builds and and runs locally, container starts and is available
• ✅ Works in tutorial account/setup
• ✅ Worker deploys & builds successfully
• ❌ Container runtime unavailable in production

[alexander-zuev]

Container image runs locally

In prod they fail with the error described above and never start

[mikenomitch]

Hey @alexander-zuev,

I think this might be an issue where "max_instances" being unset accidentally defaults it to 0, and then you "hit your limit" before any are deployed.

I'd try with a max_instances of 10 and see if it resolves. - Note that it'll take a little bit on first deploy to prep your containers, so wait until you don't see "provisioning" in the Container index of the UI.

We've got a fix coming for max_instances and nicer UI + errors here if you do hit the limit.

If that's not it, let me know and we can debug more!

[alexander-zuev]

Hey @mikenomitch! Thanks for the Discord response about container issues - I tried your suggestion to add max_instances but unfortunately it didn't resolve the problem.

Quick Context: Building https://claudeportable.dev, which allows users to run Claude Code from Slack. The full flow works locally but fails in production due to container startup issues.

Core Issue: Intermittent container provisioning failures with the error:
Error: There is no container instance that can be provided to this Durable Object, try again later

What I've Tested:

1. ✅ Created minimal test container (297MB FastAPI vs 1.26GB main container) - same issue
2. ✅ Added max_instances: 5 per your suggestion - no improvement
3. ✅ Implemented proper manual start pattern with lifecycle hooks
4. ✅ Verified containers work locally with Docker

IDs:

• basic-test-container: a0371fca-5bab-4bc6-8923-787d6282f468 (deleted)
• claude-code-container: a03db2e8-6c9f-4183-b8b1-06727249bf29

Error summary

Experiencing intermittent container startup failures when deploying a test container to Cloudflare Workers. The container fails with "There is no container instance that can be
provided to this Durable Object" errors, but occasionally succeeds after multiple retries.

Issue Description

Error Pattern (from latest logs):

• Container fails to start ~80% of the time
• Multiple consecutive failures, then occasionally succeeds
• When it fails: "no container instance" → "not listening on TCP" → "blockConcurrencyWhile() timeout"
• When it succeeds: Normal startup sequence and works perfectly

Key Observations:

• Issue affects both large (1.26GB) and small (297MB) containers equally
• Success rate is unpredictable - sometimes works after 10+ retries
• When successful, containers perform flawlessly
• Exit codes 137/143 suggest container termination during startup

Setup Details

Architecture

• Main Container: Claude Code SDK container (1.26GB)
• Test Container: Basic FastAPI container (297MB)
• Both use Cloudflare Container API with Durable Objects
• Manual start pattern with per-instance environment variables

Configuration (wrangler.jsonc)

  {
    "containers": [
      {
        "name": "claude-code-container",
        "class_name": "ClaudeCodeContainerDO",
        "image": "./Dockerfile",
        "instance_type": "basic",
        "max_instances": 1
      },
      {
        "name": "basic-test-container",
        "class_name": "BasicTestContainerDO",
        "image": "./basic_container/Dockerfile",
        "instance_type": "basic",
        "max_instances": 1
      }
    ]
  }

What We Tried

1. Created Minimal Test Container

• Built a simple FastAPI container (297MB vs 1.26GB)
• Basic endpoints: /, /health, /basic-test/{id}
• Minimal dependencies to rule out size/complexity issues

2. Implemented Manual Start Pattern

  // BasicTestContainerDO.ts
  export class BasicTestContainerDO extends Container {
    manualStart = true;
    defaultPort = 8080;
    sleepAfter = "5m";
    enableInternet = true;

    override async onStart(): Promise<void> {
      await this.ctx.storage.put('containerReady', true);
    }
  }

3. Route Implementation

  app.get('/basic-test/:id', async c => {
    const container = c.env.BASIC_TEST_CONTAINER.get(containerId);

    await container.start({
      envVars: {
        CONTAINER_TYPE: "basic-test",
        MESSAGE: `Hello from container ${id}!`,
        TEST_ID: id
      }
    });

    return await container.fetch(c.req.raw);
  });

4. Container Health Checks

• Container exposes /health endpoint
• FastAPI starts on port 8080
• Verified locally with Docker

[alexander-zuev]

UPD: Implemented retry logic which wraps .start()- still experiencing 100% failure rate

What I added:

• Exponential backoff retry mechanism (5 attempts: 500ms → 1s → 2s → 4s → 8s)
• Enhanced error logging to capture error structure
• Discovered errors have remote property indicating infrastructure-level failures

Latest production logs show consistent pattern:
1:32:49 PM - GET /test-container
Attempt 1: Failed after 4.5s - Error properties: ["stack", "message", "remote"]
Attempt 2: Failed after 4.5s - Error properties: ["stack", "message", "remote"]
Attempt 3: Failed after 4.5s - Error properties: ["stack", "message", "remote"]
Attempt 4: Failed after 4.5s - Error properties: ["stack", "message", "remote"]
Attempt 5: Failed after 4.5s - Error properties: ["stack", "message", "remote"]
Result: All retry attempts exhausted

Key findings:

1. The remote property on errors confirms this is a platform-level issue
2. Each attempt fails after ~4.5 seconds (likely internal timeout)
3. Retry logic is working correctly but cannot overcome infrastructure unavailability
4. Same behavior with both 297MB test container and 1.26GB production container

Current impact: 100% failure rate for container provisioning in production. The retry mechanism confirms containers are consistently unavailable rather than experiencing transient
failures.

Logs:

ClaudeCodeContainerDO.start - Ok @ 7/8/2025, 1:33:05 PM
  (warn) Error checking if container is ready: There is no container instance that can be provided to this Durable Object, try again later
  (warn) Error checking if container is ready: There is no container instance that can be provided to this Durable Object, try again later
✘ [ERROR]   Error: There is no container instance that can be provided to this Durable Object, try again later


✘ [ERROR]   Error: There is no container instance that can be provided to this Durable Object, try again later


Alarm @ 7/8/2025, 1:33:09 PM - Ok
  (warn) Error checking if container is ready: There is no container instance that can be provided to this Durable Object, try again later
✘ [ERROR]   Error: There is no container instance that can be provided to this Durable Object, try again later


Alarm @ 7/8/2025, 1:33:10 PM - Ok
Alarm @ 7/8/2025, 1:33:15 PM - Ok
  (warn) Error checking if container is ready: There is no container instance that can be provided to this Durable Object, try again later
  (warn) Error checking if container is ready: There is no container instance that can be provided to this Durable Object, try again later
✘ [ERROR]   Error: There is no container instance that can be provided to this Durable Object, try again later


✘ [ERROR]   Error: There is no container instance that can be provided to this Durable Object, try again later


Alarm @ 7/8/2025, 1:33:16 PM - Ok
  (warn) Error checking if container is ready: There is no container instance that can be provided to this Durable Object, try again later
  (warn) Error checking if container is ready: There is no container instance that can be provided to this Durable Object, try again later
✘ [ERROR]   Error: There is no container instance that can be provided to this Durable Object, try again later


✘ [ERROR]   Error: There is no container instance that can be provided to this Durable Object, try again later


ClaudeCodeContainerDO.start - Ok @ 7/8/2025, 1:33:14 PM
  (warn) Error checking if container is ready: There is no container instance that can be provided to this Durable Object, try again later
  (warn) Error checking if container is ready: There is no container instance that can be provided to this Durable Object, try again later
✘ [ERROR]   Error: There is no container instance that can be provided to this Durable Object, try again later


✘ [ERROR]   Error: There is no container instance that can be provided to this Durable Object, try again later


GET https://claude-code-portable.azuev.workers.dev/test-container - Ok @ 7/8/2025, 1:32:49 PM
  (log) [INFO] App - Test container route called
  (log) [INFO] Domain - User Claude config created | {"userId":"test-user-123","repoUrl":"https://github.com/alexander-zuev/supabase-mcp-server.git"}
  (log) [INFO] App - Starting container for test
  (log) [INFO] ContainerService:Domain - Starting container | {"userId":"test-user-123","repoUrl":"https://github.com/alexander-zuev/supabase-mcp-server.git"}
  (log) [INFO] ContainerService:Domain - Calling containerDO.start() with env vars | {"userId":"test-user-123","hasAnthropicKey":true,"hasGithubToken":true,"repoUrl":"https://github.com/alexander-zuev/supabase-mcp-server.git"}
  (log) [INFO] ContainerService:Domain - Attempting container start | {"userId":"test-user-123","attempt":1,"maxRetries":5}
  (error) [ERROR] ContainerService:Domain - containerDO.start() failed - detailed error analysis | {"userId":"test-user-123","attempt":1,"maxRetries":5,"errorType":"Error","errorMessage":"there is no container instance that can be provided to this durable object","errorString":"Error: there is no container instance that can be provided to this durable object","errorProperties":["stack","message","remote"],"isRetryable":true}
  (warn) [WARN] ContainerService:Domain - Retrying container start after delay | {"userId":"test-user-123","attempt":1,"nextAttempt":2,"delayMs":539}
  (log) [INFO] ContainerService:Domain - Attempting container start | {"userId":"test-user-123","attempt":2,"maxRetries":5}
  (error) [ERROR] ContainerService:Domain - containerDO.start() failed - detailed error analysis | {"userId":"test-user-123","attempt":2,"maxRetries":5,"errorType":"Error","errorMessage":"there is no container instance that can be provided to this durable object","errorString":"Error: there is no container instance that can be provided to this durable object","errorProperties":["stack","message","remote"],"isRetryable":true}
  (warn) [WARN] ContainerService:Domain - Retrying container start after delay | {"userId":"test-user-123","attempt":2,"nextAttempt":3,"delayMs":1014}
  (log) [INFO] ContainerService:Domain - Attempting container start | {"userId":"test-user-123","attempt":3,"maxRetries":5}
  (error) [ERROR] ContainerService:Domain - containerDO.start() failed - detailed error analysis | {"userId":"test-user-123","attempt":3,"maxRetries":5,"errorType":"Error","errorMessage":"there is no container instance that can be provided to this durable object","errorString":"Error: there is no container instance that can be provided to this durable object","errorProperties":["stack","message","remote"],"isRetryable":true}
  (warn) [WARN] ContainerService:Domain - Retrying container start after delay | {"userId":"test-user-123","attempt":3,"nextAttempt":4,"delayMs":2112}
  (log) [INFO] ContainerService:Domain - Attempting container start | {"userId":"test-user-123","attempt":4,"maxRetries":5}
  (error) [ERROR] ContainerService:Domain - containerDO.start() failed - detailed error analysis | {"userId":"test-user-123","attempt":4,"maxRetries":5,"errorType":"Error","errorMessage":"there is no container instance that can be provided to this durable object","errorString":"Error: there is no container instance that can be provided to this durable object","errorProperties":["stack","message","remote"],"isRetryable":true}
  (warn) [WARN] ContainerService:Domain - Retrying container start after delay | {"userId":"test-user-123","attempt":4,"nextAttempt":5,"delayMs":4054}
  (log) [INFO] ContainerService:Domain - Attempting container start | {"userId":"test-user-123","attempt":5,"maxRetries":5}
  (error) [ERROR] ContainerService:Domain - containerDO.start() failed - detailed error analysis | {"userId":"test-user-123","attempt":5,"maxRetries":5,"errorType":"Error","errorMessage":"there is no container instance that can be provided to this durable object","errorString":"Error: there is no container instance that can be provided to this durable object","errorProperties":["stack","message","remote"],"isRetryable":true}
  (error) [ERROR] ContainerService:Domain - All retry attempts exhausted | {"userId":"test-user-123","finalAttempt":5,"error":"Error: there is no container instance that can be provided to this durable object"}
  (error) [ERROR] ContainerService:Domain - Failed to start container | {"userId":"test-user-123","error":"Error: there is no container instance that can be provided to this durable object","stack":"Error: there is no container instance that can be provided to this durable object\n    at async _ContainerService.startContainerWithRetry (index.js:16876:29)\n    at async _ContainerService.startContainer (index.js:16938:7)\n    at async Array.<anonymous> (index.js:18880:5)"}
  (error) [ERROR] App - Test container failed | {"error":"Error: there is no container instance that can be provided to this durable object"}
Alarm @ 7/8/2025, 1:33:17 PM - Ok
  (warn) Error checking if container is ready: There is no container instance that can be provided to this Durable Object, try again later
✘ [ERROR]   Error: There is no container instance that can be provided to this Durable Object, try again later


Alarm @ 7/8/2025, 1:33:18 PM - Ok

[alexander-zuev]

UPD2: I've tried trimming down docker image from 1200MB to 700MB but it didn't solve the issue. Not sure if image size plays a role (but a smaller test container was eventually run). I am not sure I can cut size any further though

[mikenomitch]

@alexander-zuev thanks for all the updates and the app ID. We'll take a look and see if there's something we can learn in our logs.

Really appreciate the details!

[alexander-zuev]

Hey, @mikenomitch - just checking if there is any workaround / news on the above that I could try to resolve this issue? Really appreciate your help!