Preview

This feature is subject to the "Pre-GA Offerings Terms" in the General Service Terms section of theService Specific Terms. Pre-GA features are available "as is" and might have limited support. For more information, see thelaunch stage descriptions.

This document provides an overview of metrics for Cloud Storage FUSE and how tocollect, export, and analyze metrics data.

Cloud Storage provides in-depth metrics for Cloud Storage FUSE when you useCloud Storage FUSE to mount your buckets to Cloud Storage as file systems.Metrics can help you continuously monitor performance, identify potentialbottlenecks that can impact throughput and latency, and address issues quicklyas they occur.

For diagnostic purposes, you can collect metrics in the background when you useCloud Storage FUSE. Cloud Storage offers two ways for collecting, exporting,and analyzing metrics:

• Cloud Monitoring exporter: metrics are collected and exporteddirectly from Cloud Storage FUSE to Cloud Monitoring. Exported metrics canthen be viewed in the Google Cloud console.

• Prometheus port: metrics are exported in the Prometheus metricstext-based format at a specified Prometheus port such ashttp://Host:prometheus-port/metricsusing the Prometheus open-source monitoring and alerting toolkit.Metrics can then be viewed as an output using a client URL such ascurl http://host:port/metrics

Note: If you're using Google Kubernetes Engine (GKE), metrics are enabled by defaultstarting in GKE version 1.33.0-gke.2248000. For more information,seeCloud Storage FUSE metrics in GKE.

Pricing

Cloud Storage FUSE metrics are charged based on your usage. For moreinformation about pricing, refer toGoogle Cloud Observability pricing.

Use the Cloud Monitoring exporter

To use the Cloud Monitoring exporter to export, collect, and analyze yourmetrics from Cloud Storage FUSE to Cloud Monitoring, you must complete thesteps outlined in the following sections.

Get required roles

To get the permissions that you need to use the Cloud Monitoring exporter,grant theMonitoring Metric Writer (roles/monitoring.metricWriter)IAM role on the project you want to collect metrics for.

This predefined role contains the permissions required to use theCloud Monitoring exporter. For more information about the MonitoringMetric Writer role, seeIAM documentation aboutroles/monitoring.metricWriter.

For instructions on granting roles for projects, seeManage access to projects.

Before you begin

1. Enable theCloud Monitoring API.

2. Enable the Cloud Storage API.

   Go to Cloud Storage API

Set up the Cloud Monitoring exporter

To set up the Cloud Monitoring exporter to collect, export, and analyzeCloud Storage FUSE metrics, specify an integer value in seconds using themetrics:cloud-metrics-export-interval-secs field using aCloud Storage FUSE configuration file or the--cloud-metrics-export-interval-secs optionusing the Cloud Storage FUSE CLI.

View metrics using the Google Cloud console

To view Cloud Storage FUSE metrics using the Google Cloud console, use the followinginstructions:

1. In the Google Cloud console, go to theMetrics Explorer page.

   Goto Metrics Explorer

2. Select the project you want to view metrics for.

3. From theMetric drop-down menu, clickSelect a metric.

4. SelectVM Instance, then clickCustom and enter the name of themetric you want to view. To see which metrics Cloud Storageoffers for Cloud Storage FUSE, seeAvailable metrics.

Use Prometheus to export, collect, and analyze metrics

The following sections describe how to configure and use Prometheus toexport, collect, and analyze Cloud Storage FUSE metrics.

Configure Prometheus

To configure Prometheus, follow the instructions inConfiguring Prometheus for the Ops Agent.

Use Prometheus

1. Specify the HTTP port you want to export metrics to using either theprometheus-port field in aCloud Storage FUSE configuration file or--prometheus-portgcsfuse option using thegcsfuse CLI.

2. Validate the Prometheus metrics endpoint using the following command:

   curlhttps://localhost:8080/metrics

   If successful, the command returns an output which contains text similar tothe following:

   # HELP file_cache_read_bytes_count The cumulative number of bytes read from file cache along with read type - Sequential/Random# TYPE file_cache_read_bytes_count counterfile_cache_read_bytes_count{read_type="Random"}0file_cache_read_bytes_count{read_type="Sequential"}80# HELP file_cache_read_count Specifies the number of read requests made via file cache along with type - Sequential/Random and cache hit - true/false# TYPE file_cache_read_count counterfile_cache_read_count{cache_hit="false",read_type="Random"}215file_cache_read_count{cache_hit="false",read_type="Sequential"}5# HELP file_cache_read_latencies The cumulative distribution of the file cache read latencies along with cache hit - true/false# TYPE file_cache_read_latencies histogramfile_cache_read_latencies_bucket{cache_hit="false",le="1"}215file_cache_read_latencies_bucket{cache_hit="false",le="2"}216file_cache_read_latencies_bucket{cache_hit="false",le="3"}216file_cache_read_latencies_bucket{cache_hit="false",le="4"}216file_cache_read_latencies_bucket{cache_hit="false",le="5"}216...file_cache_read_latencies_sum{cache_hit="false"}483.62783500000023file_cache_read_latencies_count{cache_hit="false"}220# HELP fs_ops_count The cumulative number of ops processed by the file system.# TYPE fs_ops_count counterfs_ops_count{fs_op="FlushFile"}9fs_ops_count{fs_op="GetInodeAttributes"}91fs_ops_count{fs_op="LookUpInode"}584fs_ops_count{fs_op="OpenDir"}122fs_ops_count{fs_op="OpenFile"}9fs_ops_count{fs_op="ReadDir"}184fs_ops_count{fs_op="ReadFile"}220fs_ops_count{fs_op="ReleaseDirHandle"}122fs_ops_count{fs_op="ReleaseFileHandle"}9fs_ops_count{fs_op="StatFS"}10# HELP fs_ops_error_count The cumulative number of errors generated by file system operations# TYPE fs_ops_error_count counterfs_ops_error_count{fs_error_category="NOT_IMPLEMENTED",fs_op="GetXattr"}1fs_ops_error_count{fs_error_category="NOT_IMPLEMENTED",fs_op="ListXattr"}1fs_ops_error_count{fs_error_category="INTERRUPT_ERROR",fs_op="LookUpInode"}58fs_ops_error_count{fs_error_category="NO_FILE_OR_DIR",fs_op="LookUpInode"}6# HELP fs_ops_latency The cumulative distribution of file system operation latencies# TYPE fs_ops_latency histogramfs_ops_latency_bucket{fs_op="FlushFile",le="1"}9fs_ops_latency_bucket{fs_op="FlushFile",le="2"}9fs_ops_latency_bucket{fs_op="FlushFile",le="3"}9fs_ops_latency_bucket{fs_op="FlushFile",le="4"}9fs_ops_latency_bucket{fs_op="FlushFile",le="5"}9...fs_ops_latency_sum{fs_op="FlushFile"}0.28800000000000003fs_ops_latency_count{fs_op="FlushFile"}9# HELP gcs_download_bytes_count The cumulative number of bytes downloaded from GCS along with type - Sequential/Random# TYPE gcs_download_bytes_count countergcs_download_bytes_count{read_type="Sequential"}2.0971528e+08# HELP gcs_read_count Specifies the number of gcs reads made along with type - Sequential/Random# TYPE gcs_read_count countergcs_read_count{read_type="Sequential"}5```

Available metrics

Cloud Storage FUSE provides the following metrics.

File system metrics

File system metrics provide insight into the operations performed on themounted bucket such as latency and error rates.

Metric name	Metric description
fs/ops_count	The number of operations processed by the file system. This metric allows grouping byop_type to get counts for individual operations.
fs/ops_error_count	The number of errors generated by file system operations. This metric can be grouped byop_type anderror_category. Each error is mapped to anerror_category in a many-to-one relationship.
fs/ops_latency	The cumulative distribution of file system operation latencies. This metric can be grouped byop_type.

Cloud Storage metrics

Cloud Storage metrics provide insight into the interactions betweenCloud Storage FUSE and Cloud Storage.

Metric name	Metric description
gcs/download_bytes_count	The number of bytes downloaded from Cloud Storage along with the read type. The read type specifies sequential, random, or parallel reads.
gcs/read_bytes_count	The number of bytes read from Cloud Storage objects, which is different fromdownload_bytes_count. For example, more data is fetched from Cloud Storage than is actually used by the application.
gcs/read_count	The number of reads from Cloud Storage and read type, either sequential, random, or parallel.
gcs/reader_count	The number of Cloud Storage object readers that are opened or closed. This metric's data is grouped by the input/output (I/O) method type such asopened orclosed.
gcs/request_count	The number of processed Cloud Storage requests.
gcs/request_latencies	The total distribution of Cloud Storage request latencies.
gcs/retry_count	The number of retry requests made to Cloud Storage. This metric can be grouped byretry_error_category that specifies the category of the error that triggered a retry.

File cache metrics

File cache metrics provide insight into the performance and behavior ofCloud Storage FUSE file caching operations.

Metric name	Metric description
file_cache/read_bytes_count	The number of bytes read from the file cache and the read type, either sequential or random.
file_cache/read_latencies	The total distribution of the file cache read latencies and the cache hit, either true or false.
file_cache/read_count	The number of read requests made using the file cache, the type of request such as sequential or random, and the cache hit, either true or false.

What's next

• Learn how you canimprove Cloud Storage FUSE performance.

• Forward yourCloud Storage FUSE logs to Cloud Logging.