Skip to main content
Version: Config V1 (legacy)

Monitoring

The ConfigCat Proxy provides diagnostic data via its /status and a Prometheus-compatible /metrics endpoint. These endpoints are served on a specific port, so you can separate them from the public HTTP communication.

The following diagnostics related configuration options are available:

OptionDefaultDescription
diag:
enabled: <true|false>
trueTurns the diagnostics HTTP server on/off.
diag:
port: 8051
8051The port used by the diagnostics HTTP server.
diag:
status:
enabled: <true|false>
trueTurns the hosting of the status endpoint on the diagnostics HTTP server on/off.
diag:
metrics:
enabled: <true|false>
trueTurns the Prometheus metrics on the diagnostics HTTP server on/off.

Status Endpoint

The Proxy provides status information (health check) about its components on the following endpoint:

GETOPTIONS/status

The Proxy regularly checks whether the underlying SDKs can communicate with their configured source and with the cache. This endpoint returns the actual state of these checks.

Responses:

  • 200: The status returned successfully.
  • 204: In response to an OPTIONS request.

Example Response:

{
"status": "healthy",
"sdks": {
"my_sdk": {
"key": "****************************************hwTYg",
"mode": "online",
"source": {
"type": "remote",
"status": "healthy",
"records": [
"Mon, 29 May 2023 16:36:40 UTC: [ok] config fetched"
]
}
},
"another_sdk": {
"key": "****************************************ovVnQ",
"mode": "offline",
"source": {
"type": "cache",
"status": "healthy",
"records": [
"Mon, 29 May 2023 16:36:40 UTC: [ok] reload from cache succeeded",
"Mon, 29 May 2023 16:36:45 UTC: [ok] config from cache not modified"
]
}
}
},
"cache": {
"status": "healthy",
"records": [
"Mon, 29 May 2023 16:36:40 UTC: [ok] cache read succeeded",
"Mon, 29 May 2023 16:36:40 UTC: [ok] cache write succeeded",
"Mon, 29 May 2023 16:36:40 UTC: [ok] cache read succeeded",
"Mon, 29 May 2023 16:36:45 UTC: [ok] cache read succeeded"
]
}
}

Details:

If everything is operational, each status node shows the value healthy. If a component encounters a failure, it'll put an error to its records collection. If a component's last two records are errors, its status will switch to degraded. If a component becomes operational again it'll put an [ok] to the records and will switch back to healthy.

If an SDK couldn't initialize itself neither from an external cache nor from the ConfigCat CDN, its status will be down. It means, this SDK is not able to accept evaluation requests because it doesn't have a valid config JSON to work with.

If an SDK was able to initialize from its configured source, but its last two attempts to refresh has been failed (either from cache or from the ConfigCat CDN), it will become degraded because each refresh attempt will put an error to its records collection. This means, it's still able to evaluate feature flags, but it might work on a stale config JSON.

The root status is healthy if all of the SDKs are healthy. If any of the SDKs become degraded or down, the root will also switch to degraded (or down if each SDK is down).

note

You can enable the status endpoint on the main HTTP port (default: 8050) with the HTTP configuration options.

Prometheus Metrics

You can set up the Proxy to export metrics about its internal state in Prometheus format. These metrics are served via the /metrics endpoint.

The following metrics are exported:

NameTypeDescription

configcat_http_request_duration_seconds

Histogram

Histogram of Proxy HTTP response time in seconds.

Tags:

  • route: The request's URL path.
  • method: The request's HTTP method.
  • status: The response's HTTP status.

configcat_grpc_rpc_duration_seconds

Histogram

Histogram of RPC call latency in seconds.

Tags:

  • method: The RPCs name.
  • code: The RPCs response code.

configcat_sdk_http_request_duration_seconds

Histogram

Histogram of ConfigCat CDN HTTP response time in seconds.

Tags:

  • sdk: The SDK's identifier that initiated the request.
  • route: The request's URL path.
  • status: The response's HTTP status.

configcat_stream_connections

Gauge

Number of active client connections per stream.

Tags:

  • sdk: The SDK's identifier that handles the connection.
  • type: sse or grpc.
  • flag: The streamed feature flag's key.

configcat_stream_msg_sent_total

Counter

Total number of all messages sent with streaming.

Tags:

  • sdk: The related SDK's identifier.
  • type: sse or grpc.
  • flag: The evaluated feature flag's key.
info

The Proxy also exports metrics about the Go environment, e.g., go_goroutines or go_memstats_alloc_bytes, and process-related stats, e.g., process_cpu_seconds_total.

To integrate with Prometheus, put the following scrape config—that points to the Proxy—into your Prometheus configuration:

scrape_configs:
- job_name: configcat_proxy
metrics_path: /metrics
static_configs:
- targets:
- <proxy-host>:8051