For the complete documentation index, see llms.txt. Markdown versions of all docs pages are available by appending .md to any docs URL.
agctl standalone analytics
Reference for the agctl standalone analytics command.
Summarize request counts, tokens, cost, and latency
Synopsis
Summarize request counts, tokens, cost, and latency from the request log database.
–by groups results by provider, model, response-model, status, ttft, tokens-per-second, or any logged attribute such as agentgateway.user. ttft and tokens-per-second group into histogram buckets labelled by their lower bound. –interval breaks results down over time.
agctl standalone analytics [flags]Examples
agctl standalone analytics
agctl standalone analytics --by provider,model --since 168h
agctl standalone analytics --by agentgateway.user --provider openai
agctl standalone analytics --by model --since 6h --interval 1h
agctl standalone analytics --by model,ttftOptions
--by strings Group by provider, model, response-model, status, ttft, tokens-per-second, or a log attribute (repeatable)
-h, --help help for analytics
--interval duration Break results down into buckets of this duration
--model strings Filter by requested model (repeatable)
-o, --output string Output format: table, yaml, or json (default "table")
--provider strings Filter by LLM provider (repeatable)
--response-model strings Filter by response model (repeatable)
--since duration Summarize requests newer than a relative duration (default 24h0m0s)
--status int64Slice Filter by HTTP status (repeatable)
--trace-id string Filter by trace IDOptions inherited from parent commands
--address string Standalone agentgateway admin address
--gateway string Gateway name from the config file
-k, --kubeconfig string kubeconfigSEE ALSO
- agctl standalone - Manage a standalone agentgateway
Was this page helpful?
