Dear Alibaba Cloud customer,
To enhance the long-term stability and reliability of Application Real-Time Monitoring Service (ARMS), we are migrating the application configuration push service for our self-developed agents.
The ARMS team has completed the migration of configuration data to the SLS configuration management service. We have also created a phased compatibility policy for different agent versions to minimize the impact on your business.
Review the following details to assess the impact and prepare for this change.
Background
ARMS self-developed agents (Java, Go, Python, etc.) retrieve application configurations through ACM. Since ACM has been decommissioned, ARMS will exclusively use the SLS configuration management service as the new configuration push channel.
To ensure both stability and a smooth transition, we have optimized the agent logic by introducing an intelligent fallback mechanism.
Phased compatibility policy
Agent version | Timeline | Configuration service |
| On or before March 30, 2027 | Pulls configurations from the ACM configuration management service. |
From March 30, 2027 to June 30, 2027 | Pulls configurations from the ACM configuration management service, but without an SLA. Configuration pulls may fail. We recommend that you upgrade to the latest agent version as soon as possible. | |
After June 30, 2027 | Agents will be unable to pull configurations from the ACM configuration management service. The custom application configuration feature will become unavailable, and agents will revert to default settings. This will affect features such as data collection volume and sample rate configuration. We strongly recommend upgrading your agent immediately to prevent disruptions to your production services. | |
| On or before March 30, 2027 | Pulls configurations from the SLS configuration management service by default. If the pull fails due to issues such as network errors or timeouts, the agent automatically falls back to ACM. |
From March 30, 2027 to June 30, 2027 | Pulls configurations from the SLS configuration management service by default. If the pull fails, the agent falls back to ACM, but without an SLA. Configuration pulls from ACM may fail. | |
After June 30, 2027 | Pulls configurations exclusively from the SLS configuration management service. |
This policy provides a sufficient migration window for a gradual transition to the new architecture.
Affected agent versions
This policy affects the following self-developed agents:
Java agent: Support for the SLS configuration management service begins with version 4.7.0.
Go agent: Support for the SLS configuration management service begins with version 2.3.0.
Python agent: Support for the SLS configuration management service begins with version 2.3.0.
An agent's version is determined by its build or release date, not its deployment date.
Potential impacts
1. Increased configuration delay
Due to the polling mechanism of the SLS configuration management service, the time for configuration changes to take effect will increase from a few seconds to approximately 10 seconds.
If your application relies on frequent dynamic configuration updates, assess whether this delay is acceptable for your use case.
2. New network connectivity requirements
The SLS configuration management service uses different domain names and ports than ACM. If your application runs in a VPC, a container, or a restricted network environment, you must ensure that it can access the internal endpoint for the SLS configuration management service in the corresponding region.
For the internal domain names and endpoints of the SLS configuration management service in each region, see the following documentation:
To access the service over an internal network: VPC-specific internal domain names.
To access the service over the public network: Endpoints.
After June 30, 2027, if agents cannot access the SLS configuration management service, they will be unable to retrieve dynamic configurations. This may cause features such as monitoring, sample rate settings, and feature toggles to malfunction.
Long-term recommendations
Upgrade your agents to a version that supports the SLS configuration management service.
Review your network policies to ensure your application instances can access the internal or public endpoint of the SLS configuration management service in your region.