Documentation for all metric collectors in the TelemetryFlow Agent.
The Kubernetes collector uses four data sources:
| Source | Used by |
|---|---|
| Kubernetes API server | Nodes, Pods, Deployments, Workloads, Storage, HPA, PDB, Events, Resource Counts, Pod Logs |
| metrics-server (MetricsV1beta1) | Node and pod/container CPU+memory usage |
Kubelet /stats/summary (proxied) |
Node CPU-ns, memory working set, filesystem, imageFs, network; Container ephemeral storage + working set |
cAdvisor /metrics/cadvisor (proxied) |
Container CPU throttle seconds (container_cpu_cfs_throttled_seconds_total) |
| Document | Description |
|---|---|
| KUBERNETES-NODES.md | Node capacity, allocatable, usage, filesystem, network |
| KUBERNETES-PODS.md | Pod phase, container resources, ephemeral storage, working set, CPU throttle, termination |
| KUBERNETES-DEPLOYMENTS.md | Deployment replica counts and rollout conditions |
| KUBERNETES-WORKLOADS.md | StatefulSets, DaemonSets, ReplicaSets, Jobs, CronJobs |
| KUBERNETES-STORAGE.md | PersistentVolumes and PersistentVolumeClaims |
| KUBERNETES-NETWORK.md | Namespace-level network I/O from Kubelet summary |
| KUBERNETES-HPA.md | HorizontalPodAutoscaler replicas and conditions |
| KUBERNETES-PDB.md | PodDisruptionBudget health and disruption budget |
| KUBERNETES-EVENTS.md | Kubernetes events and aggregate event counts |
| KUBERNETES-RESOURCE-COUNTS.md | Secrets, ConfigMaps, Ingresses per namespace |
| KUBERNETES-POD-LOGS.md | Pod container log collection |
| Document | Description |
|---|---|
| NODE-EXPORTER.md | Full node_exporter equivalent: CPU, memory, disk I/O, filesystem, network, load, thermal, textfile |
| SYSTEM.md | Lightweight host metrics + rich SystemInfo for agent heartbeat |
| Document | Description |
|---|---|
| DOCKER.md | Docker Engine API: container CPU, memory, network, disk I/O, PIDs |
| CADVISOR.md | cAdvisor Prometheus scraper: container_* and machine_* metrics |
| Document | Description |
|---|---|
| EBPF.md | eBPF: syscalls, TCP/UDP, file I/O, scheduler, memory faults, TCP state, Cilium Hubble |
| Collector | Source | Description |
|---|---|---|
| Amazon Aurora | AWS SDK (CloudWatch GetMetricData, RDS DescribeDBClusters/Instances, Performance Insights GetResourceMetrics) | 60+ CloudWatch metrics across storage, replication, cache, latency, transactions, availability; automatic topology discovery; PI-enabled instance collection; multi-cluster with IAM role assumption; push-based export to TFO Platform |
| MySQL/MariaDB/Percona | Direct database connection (Go SQL driver) | Global status/variables, InnoDB engine parsing, replication (multi-source), Galera cluster, query analytics (performance_schema), schema metrics, MariaDB-specific (Aria, ColumnStore, Spider, query cache, thread pool, user stats) |
| PostgreSQL / RDS PostgreSQL | Direct database connection (pgx) | pg_stat_activity, pg_stat_database, pg_stat_bgwriter, pg_stat_statements, table stats, replication status; RDS variant with TLS and CA bundle support |
| Microsoft SQL Server | Direct database connection (go-mssqldb) | Wait stats categorization, performance counters, index usage/fragmentation, tempdb space, agent job status, query store top queries, database file I/O |
| MongoDB | MongoDB driver | Server status, replica set status, sharding detection, query profiler (system.profile), collection/document stats, connection pool metrics |
| ClickHouse | HTTP API (JSONEachRow) | System tables (queries, parts, merges, replication_queue, databases, tables), query metrics, merge performance |
| CockroachDB | Direct database connection (pgx) | SQL stats (crdb_internal.node_statement_statistics), range stats, store metrics, replication status |
| TimescaleDB | Direct database connection (pgx) | Hypertable stats (dimensions, chunks, compression), chunk age bucketing, continuous aggregate refresh lag, background job health (stuck detection, failures), data node availability |
| SQLite3 | File-based (database/sql) | Page cache hit/miss, WAL metrics, lock contention, integrity checks, table row counts and page stats |
| Collector | Source | Description |
|---|---|---|
| Redis | RESP (INFO, commandstats) | Connections, memory, keyspace hits/misses, evictions, per-DB keys, per-command stats; optional TLS |
| Valkey | RESP (INFO, commandstats) | Same coverage as Redis under db.valkey.* — Valkey speaks RESP, so the redis client is reused |
| Memcached | TCP (stats, stats slabs) | Connections, cmd counts, hit/miss ratios, bytes, items, evictions, per-slab stats |
| RabbitMQ | Management HTTP API | Cluster object/queue totals, message flow counters & rates, per-node mem/disk/FD/proc, per-queue messages/consumers/rates; queue_filter regex |
| Apache Kafka | JMX Prometheus exporter | Broker/topic counters & gauges normalized from the JMX exporter exposition under queue.kafka.*; cluster + instance labels |
| Confluent Kafka | Confluent Metrics API | Per-topic received/sent bytes & records, retained bytes, partition count via the Confluent Cloud/Platform metrics query API |
| NATS | HTTP monitoring (/varz, /jsz) | Connections, subscriptions, sent/received msgs & bytes, routes, subscription cache hit rate, JetStream streams/consumers/messages/bytes |
| Google Cloud Pub/Sub | Cloud Monitoring API | Undelivered/outstanding messages, oldest unacked age, sent/ack counts per subscription; service-account JWT (stdlib RS256, no Google SDK) |
No external client libraries: cache/queueing/messaging collectors speak their native wire protocols or management APIs directly and forward metrics through the standard OTLP pipeline.
| Prefix | Collector |
|---|---|
k8s.* |
Kubernetes collector |
node.* |
Node Exporter collector |
system.* |
System Host collector |
container.* |
Docker collector |
container_* / machine_* |
cAdvisor (original Prometheus names) |
ebpf.* |
eBPF collector |
aurora.* |
Amazon Aurora collector |
db.mysql.* |
MySQL/MariaDB collector |
db.postgresql.* |
PostgreSQL collector |
mssql.* |
Microsoft SQL Server collector |
db.mongodb.* |
MongoDB collector |
db.clickhouse.* |
ClickHouse collector |
db.cockroachdb.* |
CockroachDB collector |
db.timescaledb.* |
TimescaleDB collector |
db.sqlite3.* |
SQLite3 collector |
db.redis.* |
Redis cache collector |
db.valkey.* |
Valkey cache collector |
db.memcache.* |
Memcached cache collector |
queue.rabbitmq.* |
RabbitMQ collector |
queue.kafka.* |
Apache Kafka collector |
queue.confluent_kafka.* |
Confluent Kafka collector |
messaging.nats.* |
NATS collector |
messaging.pubsub.* |
Google Cloud Pub/Sub collector |