Operations Stuck in Camunda

Camunda 8 Topics Discussion & Questions
1

Greetings everyone,

I have been experiencing a recurring issue in Camunda and would like to know if anyone has faced a similar situation or can provide guidance on how to resolve it.

During certain operations — such as moving an instance, canceling a process, or attempting to delete a process after it has been canceled — the system gets stuck.
On the interface, the blue progress bar reaches the end but never completes the operation. When I refresh the page, the loading starts again, but the issue persists, and the action is never finalized.

From what I can observe, it seems that some background processing does not finish, and over time this begins to affect even new instances.

So far, the only way I have found to permanently resolve the issue is by clearing all volumes, including those from Elasticsearch. However, this also removes the entire process history, which is not a viable solution.

Has anyone encountered this problem before? Do you know why it happens and whether there is a proper way to fix it without losing historical data?

Thank you in advance for your support.

image
image1118×466 21.7 KB

2

Hi there! :wave:

I can see you’re experiencing a frustrating issue with operations getting stuck in Camunda Operate. This is actually a known problem that typically occurs due to synchronization issues between Zeebe, Operate, and Elasticsearch.

Classification: This is a Problem - operations are not behaving as expected.

Based on your description, this appears to be caused by one or more of the following:

Most Common Causes:

  1. Elasticsearch Index/Shard Issues - Missing or corrupted indices (like operate-process-*) can cause operations to hang indefinitely
  2. Data Inconsistency - Process instances may exist in Elasticsearch but not in Zeebe, causing operations to fail silently
  3. Importer/Exporter Pipeline Problems - The synchronization between Zeebe and Elasticsearch may be disrupted
  4. Resource Bottlenecks - Low disk space, memory, or CPU can stall the import/export pipeline

Troubleshooting Steps (Without Losing Data):

Before we proceed with more specific solutions, I need some additional information:

  1. What version of Camunda 8 are you using? (format: major.minor.patch)
  2. Are you running SaaS or Self-Managed?
  3. What’s your deployment setup? (Docker, Kubernetes, etc.)

Immediate Actions You Can Try:

  1. Check Elasticsearch Health:

    • Verify all required indices exist and are healthy
    • Look for shard allocation issues
    • Check disk space and resource usage

  2. Review Logs:

    • Check Zeebe, Operate, and Elasticsearch logs for errors around the time operations get stuck
    • Look for import/export errors or timeouts

  3. Restart Services (Ordered):

    • Restart Operate first, then Zeebe
    • This can help resynchronize the state without data loss

  4. Resource Check:

    • Ensure adequate disk space, memory, and CPU for all components

References:

Could you please provide the version and environment details so I can give you more targeted assistance? Also, if you can share any relevant log entries from when the operations get stuck, that would be very helpful! :mag:

3

1 - I am currently using CAMUNDA 8.7.12
2 - Self-Managed
3 - Docker

4

Perfect! Thank you for the details. With Camunda 8.7.12 Self-Managed on Docker, this issue is quite common and there are specific troubleshooting steps we can take without losing your historical data.

Immediate Diagnostic Steps:

1. Check Elasticsearch Health

Run these commands to check your Elasticsearch status:

# Check cluster health
curl -X GET "http://localhost:9200/_cluster/health?pretty"

# List all indices to ensure required ones exist
curl -X GET "http://localhost:9200/_cat/indices?v"

Look for:

  • Cluster status should be green or yellow (not red)
  • Required indices like operate-*, zeebe-* should be present
  • No failed shards

2. Review Container Logs

Check logs for errors in this order:

# Elasticsearch logs
docker logs <elasticsearch-container-name> --tail=100

# Zeebe logs (look for exporter errors)
docker logs <zeebe-container-name> --tail=100

# Operate logs (look for import/query failures)
docker logs <operate-container-name> --tail=100

3. Common Error Patterns to Look For:

In Operate logs:

Exception occurred when performing post import for partition
Process instance is not yet imported for incident processing
Elasticsearch exception [type=search_phase_execution_exception, reason=all shards failed]

In Zeebe logs:

Failed to export record
Elasticsearch exporter error

Quick Recovery Steps (Data-Safe):

Step 1: Restart Services (Ordered)

# Stop services
docker stop <operate-container-name>
docker stop <zeebe-container-name>

# Start services (Zeebe first, then Operate)
docker start <zeebe-container-name>
# Wait 30 seconds
docker start <operate-container-name>

Step 2: Check Resource Usage

# Check disk space
df -h

# Check container resource usage
docker stats

Ensure you have:

  • At least 10% free disk space
  • Sufficient memory for Elasticsearch (recommended: 4GB+)

Advanced Troubleshooting:

If the basic steps don’t work, we can:

  1. Check specific indices integrity
  2. Verify Zeebe-Elasticsearch synchronization
  3. Look for specific data inconsistencies

Questions for Better Diagnosis:

  1. When did this issue first start occurring? (After an update, restart, or specific operation?)
  2. Are you seeing any specific error messages in the UI or logs?
  3. How much disk space is available on your system?
  4. What’s your Docker Compose configuration? (especially Elasticsearch memory settings)

Could you run the health check commands above and share the results? Also, if you can share any relevant log entries from when operations get stuck, that would help pinpoint the exact cause.

This approach should help us identify and fix the root cause without losing your process history! :wrench:

References: