search

kyma-project / telemetry-manager

Manager for the Kyma telemetry module
https://kyma-project.io/#/telemetry-manager/user/README
Apache License 2.0
5 stars 24 forks source link

Reduce the number of manager flags to the bare minimum #1122

Open skhalash opened 3 months ago

skhalash commented 3 months ago

Description

Reduce the number of manager CLI flags to the bare minimum: images, priority classes, manager namespace, feature flags (if any), etc.

Reasons

Due to historical reasons, we have been maintaining an extensive list of CLI flags, most of which we rarely use. This makes it difficult to locate important settings such as memory and CPU limits and requests, as they are often situated far from the code where they are applied. Additionally, these flags need to be passed between multiple types, complicating the process further.

Some of those flag names don't even make sense:

(instead of max-log-pipelines, max-metric-pipelines, max-trace-pipelines)

Attachments

flag.StringVar(&logLevel, "log-level", getEnvOrDefault("APP_LOG_LEVEL", "debug"), "Log level (debug, info, warn, error, fatal)")
flag.StringVar(&certDir, "cert-dir", ".", "Webhook TLS certificate directory")
flag.StringVar(&telemetryNamespace, "manager-namespace", getEnvOrDefault("MY_POD_NAMESPACE", "default"), "Namespace of the manager")

flag.StringVar(&traceGatewayImage, "trace-collector-image", defaultOtelImage, "Image for tracing OpenTelemetry Collector")
flag.StringVar(&traceGatewayPriorityClass, "trace-collector-priority-class", "", "Priority class name for tracing OpenTelemetry Collector")
flag.StringVar(&traceGatewayCPULimit, "trace-collector-cpu-limit", "700m", "CPU limit for tracing OpenTelemetry Collector")
flag.StringVar(&traceGatewayDynamicCPULimit, "trace-collector-dynamic-cpu-limit", "500m", "Additional CPU limit for tracing OpenTelemetry Collector per TracePipeline")
flag.StringVar(&traceGatewayMemoryLimit, "trace-collector-memory-limit", "500Mi", "Memory limit for tracing OpenTelemetry Collector")
flag.StringVar(&traceGatewayDynamicMemoryLimit, "trace-collector-dynamic-memory-limit", "1500Mi", "Additional memory limit for tracing OpenTelemetry Collector per TracePipeline")
flag.StringVar(&traceGatewayCPURequest, "trace-collector-cpu-request", "100m", "CPU request for tracing OpenTelemetry Collector")
flag.StringVar(&traceGatewayDynamicCPURequest, "trace-collector-dynamic-cpu-request", "100m", "Additional CPU request for tracing OpenTelemetry Collector per TracePipeline")
flag.StringVar(&traceGatewayMemoryRequest, "trace-collector-memory-request", "32Mi", "Memory request for tracing OpenTelemetry Collector")
flag.StringVar(&traceGatewayDynamicMemoryRequest, "trace-collector-dynamic-memory-request", "0", "Additional memory request for tracing OpenTelemetry Collector per TracePipeline")
flag.IntVar(&maxTracePipelines, "trace-collector-pipelines", 3, "Maximum number of TracePipelines to be created. If 0, no limit is applied.")

flag.StringVar(&metricGatewayImage, "metric-gateway-image", defaultOtelImage, "Image for metrics OpenTelemetry Collector")
flag.StringVar(&metricGatewayPriorityClass, "metric-gateway-priority-class", "", "Priority class name for metrics OpenTelemetry Collector")
flag.StringVar(&metricGatewayCPULimit, "metric-gateway-cpu-limit", "900m", "CPU limit for metrics OpenTelemetry Collector")
flag.StringVar(&metricGatewayDynamicCPULimit, "metric-gateway-dynamic-cpu-limit", "100m", "Additional CPU limit for metrics OpenTelemetry Collector per MetricPipeline")
flag.StringVar(&metricGatewayMemoryLimit, "metric-gateway-memory-limit", "512Mi", "Memory limit for metrics OpenTelemetry Collector")
flag.StringVar(&metricGatewayDynamicMemoryLimit, "metric-gateway-dynamic-memory-limit", "512Mi", "Additional memory limit for metrics OpenTelemetry Collector per MetricPipeline")
flag.StringVar(&metricGatewayCPURequest, "metric-gateway-cpu-request", "25m", "CPU request for metrics OpenTelemetry Collector")
flag.StringVar(&metricGatewayDynamicCPURequest, "metric-gateway-dynamic-cpu-request", "0", "Additional CPU request for metrics OpenTelemetry Collector per MetricPipeline")
flag.StringVar(&metricGatewayMemoryRequest, "metric-gateway-memory-request", "32Mi", "Memory request for metrics OpenTelemetry Collector")
flag.StringVar(&metricGatewayDynamicMemoryRequest, "metric-gateway-dynamic-memory-request", "0", "Additional memory request for metrics OpenTelemetry Collector per MetricPipeline")
flag.IntVar(&maxMetricPipelines, "metric-gateway-pipelines", 3, "Maximum number of MetricPipelines to be created. If 0, no limit is applied.")

flag.StringVar(&fluentBitMemoryBufferLimit, "fluent-bit-memory-buffer-limit", "10M", "Fluent Bit memory buffer limit per log pipeline")
flag.StringVar(&fluentBitFsBufferLimit, "fluent-bit-filesystem-buffer-limit", "1G", "Fluent Bit filesystem buffer limit per log pipeline")
flag.StringVar(&deniedFilterPlugins, "fluent-bit-denied-filter-plugins", "kubernetes,rewrite_tag,multiline", "Comma separated list of denied filter plugins even if allowUnsupportedPlugins is enabled. If empty, all filter plugins are allowed.")
flag.StringVar(&fluentBitCPULimit, "fluent-bit-cpu-limit", "1", "CPU limit for tracing fluent-bit")
flag.StringVar(&fluentBitMemoryLimit, "fluent-bit-memory-limit", "1Gi", "Memory limit for fluent-bit")
flag.StringVar(&fluentBitCPURequest, "fluent-bit-cpu-request", "100m", "CPU request for fluent-bit")
flag.StringVar(&fluentBitMemoryRequest, "fluent-bit-memory-request", "50Mi", "Memory request for fluent-bit")
flag.StringVar(&fluentBitImage, "fluent-bit-image", defaultFluentBitImage, "Image for fluent-bit")
flag.StringVar(&fluentBitExporterImage, "fluent-bit-exporter-image", defaultFluentBitExporterImage, "Image for exporting fluent bit filesystem usage")
flag.StringVar(&fluentBitPriorityClassName, "fluent-bit-priority-class-name", "", "Name of the priority class of fluent bit ")

flag.BoolVar(&enableSelfMonitor, "self-monitor-enabled", false, "Enable self-monitoring of the pipelines")
flag.StringVar(&selfMonitorImage, "self-monitor-image", defaultSelfMonitorImage, "Image for self-monitor")
flag.StringVar(&selfMonitorCPULimit, "self-monitor-cpu-limit", "0.2", "CPU limit for self-monitor")
flag.StringVar(&selfMonitorCPURequest, "self-monitor-cpu-request", "0.1", "CPU request for self-monitor")
flag.StringVar(&selfMonitorMemoryLimit, "self-monitor-memory-limit", "90Mi", "Memory limit for self-monitor")
flag.StringVar(&selfMonitorMemoryRequest, "self-monitor-memory-request", "42Mi", "Memory request for self-monitor")
flag.StringVar(&selfMonitorPriorityClass, "self-monitor-priority-class", "", "Priority class name for self-monitor")

flag.StringVar(&deniedOutputPlugins, "fluent-bit-denied-output-plugins", "", "Comma separated list of denied output plugins even if allowUnsupportedPlugins is enabled. If empty, all output plugins are allowed.")
flag.IntVar(&maxLogPipelines, "fluent-bit-max-pipelines", 5, "Maximum number of LogPipelines to be created. If 0, no limit is applied.")

flag.BoolVar(&enableWebhook, "validating-webhook-enabled", false, "Create validating webhook for LogPipelines and LogParsers.")

Release Notes

github-actions[bot] commented 1 month ago

This issue has been automatically marked as stale due to the lack of recent activity. It will soon be closed if no further activity occurs. Thank you for your contributions.