Get a comprehensive snapshot of current system status: agents, real-time counters (TPS, response time, heap, CPU), active service counts, and recent alerts. Use this as the starting point for any investigation.

Perform automated multi-step performance diagnosis. Checks system counters, active services, error patterns, slow SQL, and service hotspots. Returns a structured diagnostic report with findings ranked by severity (CRITICAL/WARNING/INFO) and suggested actions.

Get historical counter values over a time range for trend analysis. Supports TPS, ElapsedTime, HeapUsed, ProcCpu, GcCount, Cpu, Mem, and more. Use latest_sec for quick recent data (e.g., last 60s), or start_time/end_time for custom ranges. Auto-selects precision: 2-second for <=2h, 5-minute for >2h. Large datasets are auto-sampled.

Get recent real-time transaction data (XLogs) produced since the given offsets. Returns decoded XLog entries with loop/index offsets for subsequent polling. Use offset 0/0 for the first call, then pass returned offsets to get newer data.

Search for transactions (XLogs) within a time range. Find slow transactions, error transactions, or filter by service name/IP/login. Returns list sorted by elapsed time descending with summary statistics. Primary tool for 'why is it slow?' investigations.

Get full detail of a specific transaction including execution profile (SQL queries with bind params, API calls, method traces). Use after identifying a problematic transaction from search_transactions or list_active_services.

Show all currently executing requests (in-flight transactions). Use when investigating hangs, slow conditions, or asking 'what is running right now'. Returns active services sorted by elapsed time descending.

Trace a distributed transaction across multiple services using GXID, or fetch multiple transactions by txid list. Essential for MSA environments to follow a request across service boundaries. Use after finding a GXID in get_transaction_detail.

Get aggregated service-level performance statistics. Shows which services are called most, have highest error rates, or consume most resources. Includes external API call summary for dependency analysis.

Analyze SQL performance within a time range. Returns SQL statements ranked by total elapsed time, showing execution counts, error counts, and average execution time. Use when investigating database-related performance issues.

Analyze errors within a time range. Returns error types ranked by frequency with error messages, stack traces, and sample transaction IDs for drill-down. Also shows per-service error rates for context.

Get real-time service-to-service call relationships. Shows which services call which other services, with call counts, error counts, and latency. Useful for understanding service dependencies and topology.

Get unique visitor (user) statistics. Supports real-time current visitor count, daily totals, or hourly breakdown. Useful for understanding traffic patterns and user activity.

Analyze request distribution by client IP address. Shows which IPs generate the most traffic. Useful for identifying heavy users, bots, or potential DDoS sources.

Analyze request distribution by browser/user-agent. Shows which clients (browsers, bots, mobile apps) generate the most traffic. Useful for understanding client demographics and detecting crawlers.

Get aggregated alert statistics within a time range. Shows which alert types fired most frequently. Use alongside get_system_overview (which shows recent real-time alerts) for historical alert pattern analysis.

[HTTP mode only] Read alert scripting configuration for a specific counter (read-only). Shows the current alert script, threshold rules, and configuration. Use get_server_info with include_counter_model to discover available counter names.

[HTTP mode only] Read server or agent configuration (read-only). Without obj_hash, returns the collector server's scouter.conf. With obj_hash, returns that specific agent's configuration. Use get_system_overview to find agent hashes.

Get Scouter collector server metadata: version, ID, connection status. Optionally includes counter model (all available performance counter definitions with display names and units). Use this to discover what counters exist before querying get_counter_trend. Counter model requires HTTP mode.

Get host-level system information: top processes (by CPU/memory) and disk usage. Requires a host agent obj_hash, or specify obj_type (e.g., 'linux') to auto-discover host agents. Use get_system_overview to find host agent hashes.

Get detailed agent runtime information: thread list with states and active services, environment variables, and socket connections. Thread list shows live thread states (unlike thread dump which captures stack traces). Use for deep agent-level investigation.

Capture a thread dump from a specific JVM agent. Use for diagnosing deadlocks, thread contention, or understanding thread activity. Optionally includes heap histogram for memory analysis. Stack traces require HTTP mode; TCP mode returns thread states only.

Get raw (non-decoded) profile steps for a transaction. Unlike get_transaction_detail which returns decoded/human-readable profile data, this returns the raw Step objects with hash IDs. Use lookup_text to resolve hash IDs to text. Useful when you need the raw step structure for programmatic analysis.

Get raw (non-decoded) XLog transaction data. Unlike search_transactions which returns decoded data, this returns raw XLog objects with hash IDs. Supports five modes: single lookup, GXID distributed trace, filtered search, pageable list, and real-time streaming. Use lookup_text to resolve hash IDs.

Resolve hash IDs to their original text values. Scouter stores SQL queries, service names, error messages, etc. as integer hashes for efficiency. Use this to get the actual text for a hash ID seen in other tool outputs.