Skip to content

context-types.md

Latest commit

 

History

History
218 lines (174 loc) · 10.4 KB

context-types.md

File metadata and controls

218 lines (174 loc) · 10.4 KB

Context Types

Complete reference for all 74 ContextType values used by the TelemetryFlow LLM module's ContextCollector. Context types determine what telemetry data is automatically gathered when sending a chat message or generating an insight.

How Context Types Work

sequenceDiagram
    participant Tool as Hermes Tool
    participant API as TFO API
    participant CC as ContextCollector
    participant CH as ClickHouse

    Tool->>API: POST /llm/chat/message<br/>{contextType: "metrics", contextId: "payments-api"}
    API->>CC: collectContext("metrics", "payments-api", timeFrom, timeTo)
    CC->>CH: Query relevant tables based on contextType
    CH-->>CC: metrics_1m data for payments-api
    CC-->>API: Injected context in prompt
    API->>API: Send to LLM with context
    API-->>Tool: AI response with data-grounded analysis
Loading

Context Type Categories

graph TD
    CTX["ContextType<br/>74 Values"]

    CORE["Core Telemetry<br/>5 types"]
    K8S["Kubernetes<br/>8 types"]
    INFRA["Infrastructure<br/>5 types"]
    NET["Network<br/>2 types"]
    UI["Dashboards & Reports<br/>2 types"]
    IAM_SEC["IAM & Security<br/>6 types"]
    TENANT["Tenancy<br/>5 types"]
    PLAT["Platform<br/>7 types"]
    ACCT["Account<br/>6 types"]
    AI["AI Intelligence<br/>4 types"]
    DB["Database Monitoring<br/>18+ types"]

    CTX --> CORE & K8S & INFRA & NET & UI & IAM_SEC & TENANT & PLAT & ACCT & AI & DB
Loading

Core Telemetry

ContextType Description ClickHouse Source ContextCollector Action
metrics Time-series metric data metrics_1m, metrics_5m, metrics_1h Queries recent metric values for the context scope
logs Application and system logs otel_logs Searches recent log entries, filtered by service/severity
traces Distributed traces otel_traces Queries trace spans, waterfall analysis
exemplars Metric-to-trace links exemplars, exemplars_1h Joins metric data with specific traces
correlations Cross-signal correlations signal_correlations_1h Pre-computed correlation analysis

Kubernetes Monitoring

ContextType Description ClickHouse Source
kubernetes-overview Cluster overview kubernetes_metrics_1h
kubernetes-clusters Cluster-level metrics kubernetes_metrics_1h
kubernetes-namespaces Namespace metrics kubernetes_metrics_1h
kubernetes-nodes Node-level metrics kubernetes_metrics_1h
kubernetes-pods Pod metrics and status kubernetes_metrics_1h
kubernetes-deployments Deployment metrics kubernetes_metrics_1h
kubernetes-pv Persistent volume metrics kubernetes_metrics_1h
kubernetes-api-server API server metrics kubernetes_metrics_1h
kubernetes-coredns CoreDNS metrics kubernetes_metrics_1h

Infrastructure Monitoring

ContextType Description ClickHouse Source
infra-overview Infrastructure overview vm_metrics_1h
infra-cpu CPU utilization vm_metrics_1h
infra-memory Memory utilization vm_metrics_1h
infra-storage Storage utilization vm_metrics_1h
infra-network Network metrics vm_metrics_1h

Network & Service Map

ContextType Description ClickHouse Source
service-map Service dependency graph service_map_metrics_1h
network-map Network traffic map network_map_traffic_1h, network_map_connection_metrics_1h

Dashboards & Reports

ContextType Description
dashboard Dashboard configuration and data
reports Generated reports

Alert Management

ContextType Description ClickHouse Source
alerts Active alert instances Alert state data
alert-rules Alert rule configurations Alert rule definitions
agents TelemetryFlow Agent status Agent registration data

Uptime & Status Pages

ContextType Description ClickHouse Source
uptime Uptime check results uptime_checks
status-page Status page configuration Status page data

IAM & Security

ContextType Description
iam IAM overview
iam-users User management
iam-roles Role definitions
iam-permissions Permission assignments
iam-matrix Role-permission matrix
iam-assignments User-role assignments

Tenancy

ContextType Description
tenancy Tenancy overview
tenancy-regions Region management
tenancy-organizations Organization management
tenancy-workspaces Workspace management
tenancy-tenants Tenant management

Platform Configuration

ContextType Description
audit Audit log entries
retention Retention policies
subscription Subscription management
api-keys API key management
notifications Notification channels
data-masking PII masking rules
system-setup System configuration
system-channels System notification channels

AI Assistant

ContextType Description
ai-assistant AI Assistant conversation context

Account Management

ContextType Description
account-profile User profile
account-security Security settings
account-sessions Active sessions
account-notifications Account notifications
account-preferences User preferences
account-organization Account organization

AI Intelligence

ContextType Description Module
anomaly-detection Detected anomalies AI Intelligence — Anomaly Detection
corrective-maintenance Corrective actions AI Intelligence — Corrective Maintenance
cost-optimization Cost recommendations AI Intelligence — Cost Optimization
predictive-maintenance Predictions AI Intelligence — Predictive Maintenance

Database Monitoring

ContextType Description Database Engine
db-monitoring-inventory Monitored database inventory All
db-monitoring-clickhouse ClickHouse metrics ClickHouse
db-monitoring-mysql MySQL metrics MySQL
db-monitoring-postgresql PostgreSQL metrics PostgreSQL
db-monitoring-mongodb-community MongoDB Community metrics MongoDB
db-monitoring-mongodb-atlas MongoDB Atlas metrics MongoDB Atlas
db-monitoring-mssql MSSQL metrics Microsoft SQL Server
db-monitoring-sqlite3 SQLite3 metrics SQLite3
db-monitoring-timescaledb TimescaleDB metrics TimescaleDB
db-monitoring-aurora Aurora metrics Amazon Aurora
db-monitoring-aws-rds-mysql AWS RDS MySQL metrics Amazon RDS MySQL
db-monitoring-aws-rds-aurora AWS RDS Aurora metrics Amazon RDS Aurora
db-monitoring-aws-rds-postgresql AWS RDS PostgreSQL metrics Amazon RDS PostgreSQL
db-monitoring-aws-dynamodb DynamoDB metrics Amazon DynamoDB
db-monitoring-cockroachdb CockroachDB metrics CockroachDB
db-monitoring-qan Query Analytics data All (QAN)

Context Collector Behavior

The ContextCollector (4,440 lines) handles each context type differently:

graph TD
    INPUT["contextType + contextId<br/>+ timeFrom + timeTo"]
    ROUTE["Route by contextType"]

    MET["metrics → Query<br/>metrics_1m/5m/1h"]
    LOG["logs → Search<br/>otel_logs"]
    TRC["traces → Query<br/>otel_traces"]
    K8["kubernetes-* → Query<br/>kubernetes_metrics_1h"]
    IF["infra-* → Query<br/>vm_metrics_1h"]
    DB["db-monitoring-* → Query<br/>db-specific tables"]
    AI["ai-intelligence → Query<br/>AI module data"]
    PL["platform → Query<br/>PostgreSQL + ClickHouse"]

    INPUT --> ROUTE
    ROUTE --> MET & LOG & TRC & K8 & IF & DB & AI & PL
Loading

For each context type, the ContextCollector:

  1. Determines which ClickHouse tables to query
  2. Applies time range filters (timeFrom → timeTo)
  3. Applies scope filters (contextId as service name, host, etc.)
  4. Gathers relevant data
  5. Formats as context for the PromptBuilder