Mimir

Production

Observability

Agent executions, retrieval quality, model usage, tool usage, latency, costs, failures, prompt evaluations, and streaming health.

Retrieval quality

82%

12 samples

Verification pass

92%

Test generation

p95 latency

1.8s

p50 420ms

Month cost

$0.00

$0.00 today

Retrieval quality

Average relevance score across recent retrievals.

82%

Average score

Higher is better

Based on 12 scored samples

Tool usage

Calls by tool across recent agent runs.

  • Hybrid Search18
  • Retrieve Context18
  • Graph Search10
  • Build Context7

Latency

Response time percentiles for agent and tool calls.

p50Median response
420ms
p95Tail latency
1.8s

Failure rates

Share of failed agent and tool invocations.

Agent failures0.0%
Tool failures0.0%

Model & tokens

Provider usage and token split.

openai
Est. 0 tokens · $0.00

Input

0

Output

0

Token counters will populate as traffic flows.

Costs

Estimated spend for model and tool usage.

Today

$0.00

This month

$0.00

Streaming

Stream starts versus cancellations.

Started

0

Cancelled

0

Prompt evaluations

Review findings and human disposition.

Reports

8

Findings

389

False positives

0

Accepted

1

Ignored

0

Test generation

Runs, confidence, and synthesis source mix.

Runs

1

Cache entries

0

Verification pass rate92%
Average confidence81%

Synthesis sources

  • Enhanced Template2
  • Llm0
  • Template Fallback0

System health

Service readiness and integration status.

  • Database
    Ok
  • API
    Ok
  • Workers
    Configured
  • GitHub
    Configured Or Demo
  • Demo mode
    On
  • Readiness
    ready