Using Anypoint Monitoring with Runtime Fabric
Anypoint Runtime Fabric includes native support for Anypoint Monitoring, which collects and exposes metrics, logs, and traces for Mule apps and API gateways deployed to Runtime Fabric.
Metrics collection is enabled by default for all plans. Log and trace collection require an Anypoint Platform Integration Advanced subscription.
Before You Begin
Verify Network Connectivity
For metrics, logs, and traces to reach Anypoint Monitoring, each node in Runtime Fabric must be able to send outbound TCP traffic on port 443 to these domains:
-
US control plane:
dias-ingestor-router.us-east-1.prod.cloudhub.io
us1.ingest.mulesoft.com -
EU control plane:
dias-ingestor-router.eu-central-1.prod-eu.msap.io
eu1.ingest.mulesoft.com
|
If you are using Runtime Fabric older than version 1.8.50, use port 5044 instead of port 443 and these domains:
|
For the full list of required endpoints, see Configure Network for Runtime Fabric.
If your cluster uses an outbound proxy, configure Runtime Fabric to route monitoring traffic through it. See Configure Proxies on Runtime Fabric.
| Starting with Runtime Fabric agent version 2.6.22, the Anypoint Monitoring sidecar supports HTTP proxy only. SOCKS5 proxy is not supported for monitoring traffic. |
Collect Metrics
Metrics collection is automatic. When a Mule app or API gateway replica is deployed to Runtime Fabric, a monitoring sidecar is deployed alongside it and begins transmitting metrics data to Anypoint Monitoring via mutual TLS encryption.
Each replica caches up to 100 MB of metrics data. If the connection to Anypoint Monitoring is interrupted, the cache typically provides at least four days of storage per replica before data is rotated. Custom metrics may consume cache faster.
These Runtime Fabric operations do not affect or restore existing metrics data:
-
Backup and restore
-
Runtime Fabric upgrade
-
OS patching of nodes
Enable and Configure Log Forwarding
Log forwarding requires an Anypoint Platform Integration Advanced subscription. When enabled, Anypoint Monitoring provides access to log data for applications deployed to Runtime Fabric.
To enable log forwarding for a Mule app:
-
In Runtime Manager, open the application and click the Logging tab.
-
Select the Enable application logs option.
-
Select the log level from the drop-down list:
-
INFO: Informative messages
-
DEBUG: Debugging messages
-
WARNING: Warning messages
-
ERROR: Error messages, such as when an exception occurs
-
FATAL: Fatal messages for when an application fails
-
-
Optionally, set the log level for a specific Java package class by selecting the log level, entering the
package.name, and pressing the Enter key. -
Click Apply Changes.
Log Storage and Throughput
Each replica stores up to 450 MB of log data on disk. When the limit is reached, older log data is rotated to make room for new entries.
The log forwarding sidecar can transmit up to 300 KB of log data per second per replica.
In some cases, multi-line log entries may not be fully transmitted. This can occur when multiple lines are split across two files during log rotation inside Runtime Fabric.
Log Forwarding Architecture
Starting with Runtime Fabric agent version 3.0.0, logs are forwarded exclusively through a sidecar container deployed alongside each application replica. This replaces the previous DaemonSet-based approach and provides better isolation and resource management.
|
If you are upgrading from Runtime Fabric agent version 2.11.x or earlier, you must complete migration steps before upgrading to version 3.0.0. See Migrate Log Forwarding Before Upgrading Runtime Fabric. |
Enable Trace Data Collection
| This feature is available only if your organization has the Anypoint Integration Advanced package or a Titanium subscription to Anypoint Platform. For more information, see the Pricing and Packaging documentation. |
Traces provide full visibility into the execution path of API calls and application transactions. When enabled, Anypoint Platform collects tracing information from Mule runtime engine in the OpenTelemetry standard, which you can export using the Anypoint Monitoring Telemetry Exporter.
Prerequisites for Traces
-
Mule runtime version 4.6 or later
-
Runtime Fabric agent version 2.6.31 or later
-
Traces can only be enabled using the Runtime Manager UI or API. The Mule Maven plugin, Anypoint CLI, and Anypoint Studio are not supported.
Enable Tracing for an App
-
In Runtime Manager, open the application.
-
Click the Monitoring tab.
-
Select the Enable tracing checkbox.
-
Click Deploy Application for a new deployment, or Apply Changes for an existing application.
After redeployment, the Runtime Manager backend begins instrumenting traces for the application.
Allocate Resources for the Monitoring Sidecar
Starting with Runtime Fabric agent version 3.0.0, all telemetry — metrics, logs, and traces — is transported through a single sidecar container deployed per application replica. Allocate cluster capacity accordingly to avoid deployment failures.
Resource requirements per replica depend on which features are enabled:
| Feature | CPU | Memory |
|---|---|---|
Metrics only |
0.05 vCPU |
90 MiB |
Metrics + logs |
Up to 30% of the app replica’s CPU limit |
Up to 150 MiB |
Metrics + logs + traces |
Up to 30% of the app replica’s CPU limit |
Up to 240 MiB |
Legacy Architecture (Runtime Fabric Agent 2.11.x and Earlier)
In Runtime Fabric agent version 2.11.x and earlier, two separate agents handled telemetry:
-
One DaemonSet per node to forward logs
-
One sidecar per replica to forward metrics
This architecture was replaced by the single-sidecar model in version 3.0.0.
Disable Anypoint Monitoring
Anypoint Monitoring is enabled by default. To disable it for a specific Mule app, pass this property at deployment time:
anypoint.platform.config.analytics.agent.enabled=false| Disable monitoring only if your team or organization has compliance or geospecific requirements that prevent sending metrics to the Anypoint control plane, and you are not contractually obligated to send metrics for usage and billing purposes. |