Shape the future of software operations and make your voice heard by taking the 2021 State of DevOps survey.

Data storage and retrieval services

The Google Cloud data services discussed on this page include those that store and provide data as a response to a request. SLIs for these services are similar to SLIs for request-response services, described in Request-response services, with a primary focus on availability and latency. Note that latency, especially when measuring database-query response time, is often a factor of how much data is being retrieved, and can vary with application workload.

You express a request-based availability SLI by using the TimeSeriesRatio structure to set up a ratio of "good" requests to total requests. You decide how to filter the metric by using its available labels to arrive at your preferred determination of "good" or "valid".

You express a request-based latency SLI by using a DistributionCut structure.

Cloud Storage

Cloud Storage is Google Cloud's worldwide, highly durable object store. Cloud Storage is available in multiple storage classes, which let you determine the appropriate cost and retrieval model for your service or use case.

For additional information, see the following:

Availability SLIs

Cloud Storage writes metric data to Cloud Monitoring using the gcs_bucket monitored-resource type and the api/request_count metric type. You can filter the data by using the response_code metric label to count "good" requests. You can also use the method metric label to measure availability for a specific API method, such as ReadObject.

You express a request-based availability SLI for reading objects from a Cloud Storage bucket by using a ratio, TimeSeriesRatio, of good requests to total requests, as shown in the following example:

"serviceLevelIndicator": {
  "requestBased": {
    "goodTotalRatio": {
      "totalServiceFilter":
        "metric.type=\"storage.googleapis.com/api/request_count\"
         resource.type=\"gcs_bucket\"
         metric.label.\"method\"=\"ReadObject\"
         resource.label.\"bucket_name\"=\"my_bucket\"",
      "goodServiceFilter":
        "metric.type=\"storage.googleapis.com/api/request_count\"
         resource.type=\"gcs_bucket\"
         metric.label.\"method\"=\"ReadObject\"
         resource.label.\"bucket_name\"=\"my_bucket\"
         metric.label.\"response_code\"=\"OK\"",
    }
  }
}

Latency SLIs

Cloud Storage doesn't provide a latency metric; if latency is a concern, consider instrumenting your service to collect this on the client.

Cloud Bigtable

Cloud Bigtable is a fully managed, scalable NoSQL database service for large analytical and operational workloads. Bigtable is ideal for storing very large amounts of data in a key-value store. Bigtable also supports high read-and-write throughput at low latency, offering you fast access to large amounts of data.

For additional information, see the following:

Availability SLIs

Bigtable writes metric data to Cloud Monitoring using the bigtable_table monitored-resource type and the following metric types:

You express a request-based availability SLI for reading objects from a Cloud Storage bucket by using a ratio, TimeSeriesRatio, of "bad" requests to total requests, as shown in the following example:

You can use these two metrics to express a request-based availability SLI by creating a TimeSeriesRatio structure for failed requests to total requests, as shown in the following example:

"serviceLevelIndicator": {
  "requestBased": {
    "goodTotalRatio": {
      "totalServiceFilter":
        "metric.type=\"bigtable.googleapis.com/server/request_count\"
         resource.type=\"bigtable_table\"
         resource.label.\"table\"=\"my_table\"
         resource.label.\"cluster\"=\"my_cluster\"",
      "badServiceFilter":
        "metric.type=\"bigtable.googleapis.com/server/error_count\"
         resource.type=\"bigtable_table\"
         resource.label.\"table\"=\"my_table\"
         resource.label.\"cluster\"=\"my_cluster\"",
    }
  }
}

Latency SLIs

To measure latency, Bigtable writes metric data to Cloud Monitoring using the bigtable_table monitored-resource type and the server/latencies metric type. You can filter the data by using the method metric label to measure the latencies of specific methods.

You express a request-based latency SLI by using a DistributionCut structure.

The following example SLO expects 99% of all requests to the my_table table in the my_cluster cluster to fall between 0 and 100 ms in total latency over a rolling one-hour period:

{
  "serviceLevelIndicator": {
    "requestBased": {
      "distributionCut": {
        "distributionFilter":
          "metric.type=\"bigtable.googleapis.com/server/latencies\"
           resource.type=\"bigtable_table\"
           resource.label.\"table\"=\"my_table\"
           resource.label.\"cluster\"=\"my_cluster\"",
        "range": {
          "min": 0,
          "max": 100
        }
      }
    }
  },
  "goal": 0.99,
  "rollingPeriod": "86400s",
  "displayName": "98% requests under 100 ms"
}

Cloud Spanner

Cloud Spanner is a fully managed relational database service that offers transactional consistency at global scale, schemas, SQL (ANSI 2011 with extensions), and automatic, synchronous replication for high availability.

For additional information, see the following:

Availability SLIs

Spanner writes metric data to Cloud Monitoring using the spanner_instance monitored-resource type and the query_count metric type. You can filter the data by using the status metric label to count successful and failed database queries.

You express a request-based availability SLI by creating a TimeSeriesRatio structure for "good" requests to total requests, as shown in the following example:

"serviceLevelIndicator": {
  "requestBased": {
    "goodTotalRatio": {
      "totalServiceFilter":
        "metric.type=\"spanner.googleapis.com/query_count\"
         resource.type=\"spanner_instance\"
         metric.label.\"database\"=\"my_database\"",
      "goodServiceFilter":
        "metric.type=\"spanner.googleapis.com/query_count\"
         resource.type=\"spanner_instance\"
         metric.label.\"database\"=\"my_database\"
         metric.label.\"status\"=\"ok\"",
    }
  }
}

Latency SLIs

To measure latency, Spanner writes metric data to Cloud Monitoring using the spanner_instance monitored-resource type and the api/request_latencies metric type. You can filter the data by using the method metric label to measure the latencies of specific methods. The data includes latencies not just for queries, but for other Spanner API calls.

You express a request-based latency SLI by using a DistributionCut structure. The following example SLO expects that 99% of all API requests to the my_database database fall between 0 and 100 ms in total latency over a rolling one-hour period:

{
  "serviceLevelIndicator": {
    "requestBased": {
      "distributionCut": {
        "distributionFilter":
          "metric.type=\"spanner.googleapis.com/api/request_latencies\"
           resource.type=\"spanner_instance\"
           metric.label.\"database\"=\"my_database\"",
        "range": {
          "min": 0,
          "max": 100
        }
      }
    }
  },
  "goal": 0.99,
  "rollingPeriod": "86400s",
  "displayName": "98% requests under 100 ms"
}

Datastore

Datastore is a highly scalable NoSQL database for your applications. Datastore automatically handles sharding and replication, providing you with a highly available and durable database that scales automatically to handle your applications' loads.

For additional information, see the following:

Availability SLIs

Spanner writes metric data to Cloud Monitoring using the datastore_request monitored-resource type and the api/request_count metric type. You can filter the data by using the response_code metric label to count successful and failed API calls or by using the api_method metric label to, for example, only measure document reads.

You express a request-based availability SLI by creating a TimeSeriesRatio structure for successful requests to total requests, as shown in the following example:

"serviceLevelIndicator": {
  "requestBased": {
    "goodTotalRatio": {
      "totalServiceFilter":
        "metric.type=\"datastore.googleapis.com/api/request_count\"
         resource.type=\"datastore_request\"",
      "goodServiceFilter":
        "metric.type=\"datastore.googleapis.com/api/request_count\"
         resource.type=\"datastore_request\"
         metric.label.\"response_code\"=\"success\"",
    }
  }
}

Latency SLIs

Datastore doesn't currently provide a latency metric. For the available metrics, see datastore.googleapis.com metric types.