metrics.md

Title	description	weight	alwaysopen	categories	draft
Metrics and usage statistics	Metrics API provides programmatic access to database usage and performance data	40	false	RC	true

Metrics API provides programmatic access to database usage and performance data. The metrics API shows data that similar to the data that the Redis Cloud admin console shows in the database metrics.

Metric spans and intervals

Metrics API data shows the data for the database that you request and the metric span that you request. The metric span determines the time period for the data points in the metrics API response.

Metric Span	Data duration	Data point interval	Data point collection
1minute	1 minute	10 seconds	on-demand
5minutes	5 minutes	1 minute	on-demand
1hour	1 hour	5 minutes	continuous
1day	1 day	15 minutes	continuous
1week	1 week	1 hour	continuous
1month	1 month	12 hours	continuous
1year	1 year	7 days	continuous

{{< note >}} The data point interval represents a best-effort and is not an absolute or unvarying interval. The exact interval between one data point and the next may vary and the interval values may change. {{< /note >}}

High frequency, on-demand metric spans

The 1minute and 5minutes metric spans are high frequency intervals that require on-demand collection. The API request collects the data since the last request (up to 5 minutes) only when you request it.

For example:

• At the first request for a 1minute metric span, the metrics API operation returns an empty data-points array. This is because the first request triggers the collection of high-frequency metrics data for this metric span, but does not contain any previously collected data points.
• The next 1minute request to the database within 5 minutes shows the metrics collected since the last request.
• Any break in requests to the database for more than 5 minutes stops the data collection. The next 1minute request to the database starts the data collection again.

Metric data points

A metric data point contains these properties:

Property	Description
Category	Each data point is shown in a category and with the measurement unit for that category, for example: milliseconds, megabytes, percentage
Value	The value of the data point, defined with the measurement unit of the category for the data point
Timestamp	The exact collection time for the data point in ISO-8601 date format and in the UTC timezone, for example: 2019-03-15T14:26:02Z

For example, a request for data points in the count-connections (number of connections) category returns:

   {
      "name": "count-connections",
      "data": [
        {
          "value": 7,
          "timeStamp": "2019-08-05T16:02:22Z"
        },
        {
          "value": 9,
          "timeStamp": "2019-08-05T16:02:32Z"
        },
        {
          "value": 8,
          "timeStamp": "2019-08-05T16:02:42Z"
        }
      ]
    }

Metric categories

Data points are shown divided into these categories:

Category	Description	Units	Redis on RAM	Redis on Flash	Essentials
count-connections	Number of open connections in the database	count	Y	Y	Y
count-keys	Number of keys in the database	count	Y	Y	Y
evicted-objects	Number of evicted objects	count	Y	Y	Y
expired-objects	Number of expired objects	count	Y	Y	Y
flash-used	Flash storage used by Redis-on-Flash database	bytes	Y
fragmentation-ratio-disk	Flash storage fragmentation ratio	percentage	Y
fragmentation-ratio-memory	RAM fragmentation ratio	percentage	Y	Y	Y
hits-flash	Requests found in Flash	percentage	Y
hits-ram	Requests found in RAM	percentage	Y
hits-read	Read requests found in Redis database	percentage	Y	Y	Y
hits-write	Write requests found in Redis database	percentage	Y	Y	Y
latency-average	Average latency for all requests	microseconds	Y	Y	Y
latency-average-read	Average latency for read requests	microseconds	Y	Y	Y
latency-average-write	Average latency for write requests	microseconds	Y	Y	Y
latency-other	Average latency for requests that are not read or write	microseconds	Y	Y	Y
memory-size-calculated	Total size of Redis database	bytes	Y	Y	Y
ram-flash-ratio	Ratio between RAM and Flash access	percentage	Y
ram-overhead	The percentage of RAM used for storing everything rather than pure values. I.e. keys, Redis dictionary and other key/value overheads	bytes	Y
ram-used	RAM used by Redis-on-Flash database	bytes	Y
requests-other	Number of requests that are not read or write	ops/sec	Y	Y	Y
requests-read	Number of read requests	ops/sec	Y	Y	Y
requests-total	Total number of requests	ops/sec	Y	Y	Y
requests-write	Number of write requests	ops/sec	Y	Y	Y
values-in-flash	Number of values stored in Flash	count	Y
values-in-ram	Number of values stored in RAM	count	Y

Usage example

{{% embed-code "rv/api/70-query-metrics.sh" %}}