sls_text_to_sqlLABSCO SUMMARY
The query language stays the server's problem.
SLS answers in SQL and in SPL, and CloudMonitor in its own console; knowing which to reach for is most of the work during an incident. These tools take the question and produce the query, with the time range parsed however you happened to phrase it.
What it is
The Go implementation of Alibaba Cloud's observability MCP server, giving models structured access to Log Service (SLS) and CloudMonitor (CMS). It ships as a single binary with no runtime dependencies and speaks stdio, SSE or streamable-http.
What you get
- A plain-language question turned into an SLS SQL query
- The same question expressed in SPL, for SLS's processing pipeline
- CloudMonitor 2.0 queried in natural language
- Time expressions understood several ways: relative, absolute timestamps, Grafana style, and preset keywords
- Time-series comparison with statistics, trend analysis and a difference score
- Retries with exponential backoff, a circuit breaker and graceful shutdown, with errors described in English alongside a suggested fix
Requirements
Download the binary for your platform from the project's Releases page; it needs no runtime. Credentials go in the environment as an Alibaba Cloud access key id and its secret. Start it in stdio mode for a desktop client, or set the transport in config.yaml for network mode. The earlier Python implementation remains available in the repository's v1 directory.
Setup effort
Build from source — clone the repository, then point your client at the binary
USE CASES
01
Asking a log question during an incident
State what you want to know and get an SLS SQL query built for it, with the time range as you said it.
02
Reaching for SPL when SQL is the wrong shape
Get the same question expressed in SLS's processing language instead.
03
Checking a metric without opening CloudMonitor
Query CloudMonitor 2.0 in natural language from the client you are already in.
04
Comparing this week against last
Pull both ranges and let the comparison return trend and a difference score rather than two raw series.
4 tools
Listed in project docs. Your client may show a different list.
sls_text_to_splsls_sopcms_natural_language_queryBEFORE YOU INSTALL
What to know first
Tools you get
4 tools named in the project's documentation.
Requires
- ALIBABA_CLOUD_ACCESS_KEY_ID in your environment
- ALIBABA_CLOUD_ACCESS_KEY_SECRET in your environment
INSTALLATION
Connect your client
Add the block to your client, or run the command directly.
Server setup
Configuration file
JSON
{
"mcpServers": {
"alibaba_cloud_observability": {
"command": "./bin/alibabacloud-observability-mcp-server",
"args": [
"start",
"--stdio"
],
"env": {
"ALIBABA_CLOUD_ACCESS_KEY_ID": "<YOUR_ALIBABA_CLOUD_ACCESS_KEY_ID>",
"ALIBABA_CLOUD_ACCESS_KEY_SECRET": "<YOUR_ALIBABA_CLOUD_ACCESS_KEY_SECRET>"
}
}
}
}The command line the project publishes
TERMINAL
./bin/alibabacloud-observability-mcp-server start --stdio
