Get Database Insights

This endpoint returns the Insights series of a database: service latency percentiles, read, write, engine, replication, lock-wait and Lua latency, expired entries, replication lag, throughput, bandwidth and disk size, for the whole database, one replica or every region, with each series’ latest window compared to the one before it, and how much the limit, security and Lua counters grew over the period. Latencies are in milliseconds. Available for databases on a Pro plan or with the Production Pack.

GET /redis/insights/{id}

Get Database Insights

cURL

curl --request GET \
  --url https://api.upstash.com/v2/redis/insights/{id} \
  --header 'Authorization: Basic <encoded-value>'

200

{
  "groups": [
    {
      "key": "<string>",
      "label": "<string>",
      "kind": "all"
    }
  ],
  "metrics": {},
  "comparison_window": "3h",
  "comparison": {},
  "counters": {},
  "anomalies": {}
}

Authorizations

Authorization · string · header · required

Basic authentication header of the form Basic <encoded-value>, where <encoded-value> is the base64-encoded string username:password.

Path Parameters

id · string · required

The ID of the database

Query Parameters

period · string · default: 3h

The period of the series: 1h, 3h, 12h, 1d, 3d, 7d or 30d. Defaults to 3h.

replica · string

Restrict the series to one replica, by the key of a group whose kind is replica. Other keys are rejected with 400.

by · enum<string>

Set to region to receive one series per region next to the total.

Available options: region

Response

200 - application/json

Database insights retrieved successfully

groups · object[]

Every group the database has, regardless of the scope requested

Show child attributes

metrics · object

Metric name to group key to data points. The all group is always present; region groups are added with by=region. Metrics: latency_p50, latency_p90, latency_p99, read_latency_99, write_latency_99, command_latency_99, replication_latency_999, wait_latency_99, lua_latency_p50, lua_latency_99, expired_entries (per minute), replication_lag (entries), throughput, bandwidth (bytes per second) and diskusage.

Show child attributes

comparison_window · string

Length of the latest window the comparison averages: the whole period up to three hours, the last three hours for longer periods.

Example:

"3h"

comparison · object

Metric name to group key to the mean over the latest window and over the equal window right before it. previous is absent when that earlier window has no data.

Show child attributes

counters · object

Growth of each counter over the period, for the whole database: over_size_requests (TCP and REST), over_size_record_accesses, throttled_commands (writes delayed by persistence backpressure), undeliverable (requests and connections the proxies could not hand to any replica), auth_failures (TCP and REST, including commands sent before authenticating and REST requests over the rate limit), keyspace_notifications, connections_opened, bandwidth_ingress, bandwidth_egress (bytes), rest_commands, lua_executions, lua_commands (commands run inside scripts), lua_failures, lua_timeouts and lua_memory_limits.

Show child attributes

anomalies · object

Metric name to up to three anomalies of its all series, largest first: stretches where the value stayed above both the period's median plus three standard deviations, estimated as 1.4826 times the median absolute deviation, and 30% over the median. A series with fewer than 8 points after its first positive value reports none. The median is raised to a small per-metric floor, so a nearly idle series does not report huge changes. An empty stretch at the start of the series is ignored, and a stretch longer than a third of the analyzed points is a level change, not an anomaly. Every metric is present; diskusage is never checked and always has an empty list.

Show child attributes