Metrics Reference
This page lists the metrics generated by kdb Insights Enterprise. Refer toMonitoring for details on how metrics are captured, stored and presented.
Each component generates a range of metrics. With metrics enabled on a component, kdb info, memory and CPU metrics are generated.
Database
The kdb Insights Database publishes performance metrics regarding the storage and querying of data.
See the following links for details on the metrics available:
Reliable Transport
The Reliable Transport, also known as a kdb Insights Stream, publishes status and performance metrics.
Details on the metrics available are located here.
The metrics are published by default, but can be disabled by setting the environment variable RT_EXPORT_METRICS=0.
Base infrastructure
The following sections relate the kdb Insights Enterprise base components that manage the configuration and deployment of databases and assemblies.
Package manager
Relates to the management of packages using the kxi pm command including databases, pipelines, and views created through the UI.
|
metric name |
type |
description |
|---|---|---|
|
|
counter |
Total number of requests by method, status and handler |
|
|
summary |
Size of incoming requests by handler |
|
|
summary |
Size of outgoing responses by handler |
|
|
histogram |
Count and time take to make HTTP requests |
|
|
gauge |
Number of HTTP requests in progress |
Starting with kdb Insights Enterprise version 1.17, the Package Manager (PM) supports scaling to multiple replicas.
That means all package data is stored in a centralized database outside of the package-manager service. This allows safe scaling, and data migration is handled automatically during the kxi upgrade or kxi rollback process.
To enable resiliency, add the following to your Insights values file at install or upgrade:
yaml
kxi-package-manager:
replicaCount: 2
Set the replicaCount based on your needs, as follows:
-
1for low resource usage, non-resilient. Suitable only for environments with minimal package operations or deployments. Not recommended for heavy UI usage. -
2for basic resiliency with relatively low resource overhead. Doubles the API request throughput. -
3or more for high resilience with higher resource usage and high API request throughput. Recommended for production workloads that require high availability and responsiveness.
Important
Setting replicaCount to 2 or higher is only supported in kdb Insights Enterprise 1.17+.
Information service
Each call to the Information Service API endpoint generates metrics.
|
metric name |
type |
description |
|---|---|---|
|
|
histogram |
Count and time taken to get details for a specific client ID |
|
|
counter |
Number of failed calls to get details for a specific client ID |
Operator
The kdb Insights Operator generates metrics on each attempt to interact with the SP Coordinator or Keycloak instance of each namespace.
|
metric name |
type |
description |
|---|---|---|
|
|
counter |
Number of failed requests to Keycloak |
|
|
histogram |
Count and time taken to make requests to Keycloak |
|
|
counter |
Number of failed calls to the SP Coordinator |
|
|
histogram |
Count and time taken to make requests to the SP Coordinator |
Component level metrics
License information
Each component records kdb+ license information on startup.
|
metric name |
type |
description |
|---|---|---|
|
|
guage |
kdb+ license info, including: expiry date, OS version, process cores, release date, release version |
Memory statistics
Each component records kdb+ memory statistics. Output is taken from .Q.w[] for the component container.
|
metric name |
type |
description |
|---|---|---|
|
|
gauge |
Current memory usage in bytes |
|
|
gauge |
Memory available in the heap in bytes |
|
|
gauge |
Maximum heap size so far in bytes |
|
|
gauge |
Memory limit heap in bytes |
|
|
gauge |
Mapped memory in bytes |
|
|
gauge |
Physical memory available in bytes |
|
|
gauge |
Number of symbols |
|
|
gauge |
Memory use of symbols in bytes |