XRAY: Repositories Missing from Xray Watches/Returning "Watch is not valid"

Products
JFrog_Xray
Content Type
Administration_Platform
AuthorFullName__c
Kanishk Sharma
articleNumber
000007000
FirstPublishedDate
2026-05-25T10:05:58Z
lastModifiedDate
2026-05-25

XRAY: Repositories Missing from Xray Watches/Returning "Watch is not valid"

Introduction

When using older Xray versions that have reached End of Life (such as Xray 3.78.x with Artifactory 7.90.x), you may encounter an issue where certain repositories do not appear in the Xray Watches repository list, or adding them fails with an error.

This issue has two possible causes: a broken RabbitMQ cluster in Xray HA environments or a delay in the indexing of the repository content. Both are addressed below.


Symptoms
  • One or more repositories do not appear under Watches > [select watch] > Add Repository in the Xray UI, even though other repositories are listed.
  • A PUT /xray/api/v2/watches/<WATCH-NAME> request that includes the affected repository returns "Watch is not valid" or HTTP 500.
  • The same Watch works fine when adding other repositories.
  • In HA environments, Xray logs may show 60-second timeouts on sync actions like ReloadBlockDownloadCache.

Cause 1: Broken RabbitMQ Cluster (Xray HA)

In a multi-node Xray HA setup, RabbitMQ is used for inter-node communication. If the Erlang cookies become mismatched across nodes, one or more nodes become isolated. When a Watch update is initiated, the isolated node tries to broadcast a sync action to the other nodes via RabbitMQ, times out after ~60 seconds, and returns an HTTP 500 error.

Related logs:
2026-02-16T11:43:48.689Z [jfxr ] [INFO ] [cv122rtxaa341as7] [exchange_handler:406] [main] Sending sync action 'ReloadBlockDownloadCache' message to server microservice on node [x.x.x.x](http://node2.x.x.x)

2026-02-16T11:44:48.693Z [jfxr ] [ERROR] [cv122rtxaa341as7] [watcher_handler:487] [main] Failed to reload block download cache on other nodes for watch TESTWATCH

 

How to confirm: Run the following on each Xray node:
/opt/jfrog/xray/app/third-party/rabbitmq/sbin/rabbitmqctl cluster_status

If each node reports only itself as a cluster member, the cluster is broken.


Resolution

Re-cluster RabbitMQ by following the official documentation: How to Re-cluster RabbitMQ HA for Xray HA

If RabbitMQ nodes fail to join due to hostname resolution issues, update /etc/hosts on each node to ensure proper hostname-to-IP mapping between all Xray nodes, then retry the clustering.

After re-clustering, verify the cluster is healthy:
/opt/jfrog/xray/app/third-party/rabbitmq/sbin/rabbitmqctl cluster_status
All nodes should appear as members of the same cluster. We should be able to add the problematic repository to the Watch after performing this operation.


Cause 2: Repository Not Yet Indexed

A repository only appears in the Watches repository list after Xray has successfully indexed it. If indexing is still in progress or has not been triggered, the repository will be missing from the list.

Resolution
  1. Confirm that the repository is configured for indexing at Administration > Xray Settings > Index Repositories.
  2. Wait for the initial indexing to complete. Once indexed, the repository will appear in the Watch configuration list.
  3. If the repository still does not appear after indexing, upgrade Artifactory and Xray to the latest supported versions. This behavior has been confirmed to work correctly in newer releases.

Related links: