Application & Runtime Observability
Application and data-layer monitoring without external tooling. Runtime introspection for 7 runtimes, 5 database engines, 3 web servers and load balancers, and 2 message queues — from infrastructure to application layer in one view.
Runtime Monitoring (7 Runtimes)
JVMHeap & non-heap, Young/Old GC counts and time, Thread states, Per-thread CPU, Class histogram, Native memory, File descriptors, Trigger GC
Node.jsRSS, V8 heap (used/total/limit), External & array buffers, Active handles/requests, Native/detached contexts, Trigger GC
PythonInterpreter version, RSS/PSS/VMS & shared memory, Threads, Open FDs, Disk I/O, Network connections
.NETLoaded assemblies, RSS/PSS/VMS & shared memory, Threads, Open FDs, Disk I/O, Network connections
PHP-FPMPool active/idle/total processes, Listen queue, Max children reached, Slow requests, Requests served, OPcache hit rate & memory
RubyApp type detection (Rails, Puma, Sidekiq, Unicorn), Ruby version, RSS/PSS memory, Threads, Open FDs, Disk I/O
GoGo version, Goroutine count & dump (via pprof when exposed), RSS/PSS memory, Threads, Open FDs, Disk I/O
Database Monitoring (5 Engines)
PostgreSQLConnections (active/idle/idle-in-txn), TPS, Cache hit ratio, Replication replay lag, Dead tuples & VACUUM, Checkpoints, Locks, Table sizes, Query stats
MySQL/MariaDBQueries/sec, Slow queries, InnoDB buffer pool & row locks, Threads connected/running, Replication (seconds behind master), Query stats
RedisMemory usage & peak, Ops/sec, Hit rate, Fragmentation ratio, Evicted/expired keys, Keyspace per-DB, Clients, Slow log
MongoDBOperations by type, Connections, Resident/virtual memory, WiredTiger cache, Storage sizes, Replica set members, Oplog size, Profiler
CassandraCluster & node status, Ring and gossip info, Keyspaces & tables, Compaction stats & history, Thread pools, GC stats, CQL shell
Web Server & Message Queue Monitoring
NginxRequests/sec, Active connections, Reading/writing/waiting, Accepts vs handled, Upstreams, Access & error logs, Site config with SafeApply
ApacheRequests/sec, Bytes/sec, Busy/idle workers, MPM type, Scoreboard, Loaded modules, Access & error logs
HAProxyFrontend/backend sessions & session rate, Queue, Response time, Servers up/down, Bytes in/out
RabbitMQQueue depth (ready/unacked), Publish & deliver rates, Consumers per queue, Node memory, disk & file descriptors
KafkaBrokers & controller, Topics & partitions, Under-replicated / under-min-ISR partitions, Bytes & messages in, Consumer group lag
Log Streaming & Analytics
- Fleet-wide log search with quoted phrases and AND/OR operators
- Severity filtering and severity histogram over time
- Facets by source, severity and service
- Log pattern clustering with anomaly alerts per host
- Error rate and log throughput per host
Fleet-Wide System Metrics
CPU
Usage %, cores, load averages (1/5/15), CPU temperature and throttling
Memory
Total, Used, Available, Cached, Buffers, Swap, HugePages
Disk
Usage per partition, SMART health, Disk temperature, RAID status
Network
Rx/Tx MB/s and packets/sec per interface, Errors, Drops; TCP retransmits, RSTs and listen overflows (eBPF)
I/O
Read/Write IOPS, Throughput, Read/write latency per device; latency P50/P95/P99 and utilization (eBPF)
System
Hardware inventory (CPU, board, BIOS, GPU, fans, sensors), Process list, Active connections & listening ports
Per-Host History
Time ranges: 1h, 6h, 24h, 7d
Charts: CPU, Memory, Disk and Network Rx/Tx trends
Load average gauges (1/5/15 min)
Agent-side history store, collected every 10 seconds