Monitor in-flight executions

Execution statuses

The monitoring view shows the execution ID, macro name, queued time, and execution time.

Monitor a single node

GET /rest/asyncmacros/latest/non-cluster/inflight-macro-executions

This endpoint requires System Administrator permissions and WebSudo confirmation.

Monitor a cluster

Using the Cluster Monitoring API

First, list the nodes:
 
 

GET /rest/atlassian-cluster-monitoring/cluster/nodes

Then query a specific node:
 
 

GET /rest/atlassian-cluster-monitoring/cluster/suppliers/data/com.atlassian.confluence.plugins.confluence-async-macros/inflight-macro-executions/{NODE-ID}

Using the Confluence UI

1. Go to Administration ⚙️ > General Configuration > Clustering.

2. Select a cluster node from the list.

3. Open the Inflight Executions tab to view execution IDs, macro names, queued time, and elapsed execution time.

Cancel executions

Administrators can cancel macros that are currently executing. Cancellation is useful when a runaway macro is consuming server resources or delaying users.

Cancel through the REST API

All cancellation endpoints use /rest/asyncmacros/latest/inflight/cancel. They require System Administrator permissions and WebSudo confirmation. Cancellation is immediate in the sense that the request returns after the cancellation signal has been sent.

Cancel specific executions by ID

POST /rest/asyncmacros/latest/inflight/cancel/by-execution-ids
 
 { "executionIds": ["4afd2219-f6b9-4362-af2c-218e4c1fb0af"] }

Get execution IDs from the monitoring endpoints.

Cancel all executions of selected macros

POST /rest/asyncmacros/latest/inflight/cancel/by-macro-names
 
 { "macroNames": ["blog-posts", "jira"] }

Cancel all in-flight macro executions

 POST /rest/asyncmacros/latest/inflight/cancel/all

How cancellation works

1. The REST request publishes a Confluence event.

2. The CancellationListener receives the event and calls the canceller.

3. The canceller sends Thread.interrupt() to the execution thread and sets a cancellation flag in the thread-local ExecutionCancellationContext.

4. If the macro implements isCancellable() = true and handles the signal, it stops promptly.

5. If the macro does not support cancellation, the interrupt and flag are issued, but the macro continues until it finishes or times out naturally in the background.

Cancellation effectiveness depends on whether the macro developer implemented cancellation support. The API can return successfully even when the macro continues running.

What users see when a macro is canceled

The macro output area displays:

Macro execution cancelled

Users may see the following messages in other scenarios:
 
 
 

Configure logging

1. Go to Administration ⚙️ > General Configuration > Logging and Profiling.

2. Under Default Loggers, click Add entry.

3. Add the package or class name, select the logging level, and save.

Example log messages

Macro submitted

Asynchronous execution started for space TEST ContentId ContentId{id=2523138} URL /spaces/TEST/pages/2523138/My+Page macro: jira and execution Id 847266cb-bdfa-4676-8a0a-187651899c1e

System under pressure

EWMA latency 1234ms exceeded threshold 1000ms and queue size (20) > pool size (15); running task in caller thread.

Queue no longer full but latency remains high

 Queue is no longer full, but EWMA latency is exceeded.

System returned to normal

Queue is not full & EWMA latency is within threshold. System is back to normal.

Review statistics and analytics

A scheduled job publishes execution statistics as Confluence analytics events.

Schedule: Daily at midnight, 0 0 0 * * ?

Run the statistics job manually

1. Go to Administration ⚙️ > General Configuration > Scheduled Jobs (or open /admin/scheduledjobs/viewscheduledjobs.action).

2. Locate Async Macro Statistic Job and click Run.

Each analytics event contains the following metrics for each macro and node:
  
 

Statistics are retained in memory for two days by default. Change atlassian.async.macro.execution.statistic.retention.days to adjust retention. Node identifiers in analytics events are anonymized using hashes.

Related content

For details about implementing asynchronous macro support in a macro plugin, see the corresponding developer documentation.